Fuelling Drug Discovery Against Mycobacterium Tuberculosis

More Medicines for Tuberculosis: Fuelling Drug Discovery against Mycobacterium tuberculosis

THÈSE NO 8906 (2018)

PRÉSENTÉE LE 5 OCTOBRE 2018 À LA FACULTÉ DES SCIENCES DE LA VIE UNITÉ DU PROF. COLE PROGRAMME DOCTORAL EN BIOTECHNOLOGIE ET GÉNIE BIOLOGIQUE

ÉCOLE POLYTECHNIQUE FÉDÉRALE DE LAUSANNE

POUR L'OBTENTION DU GRADE DE DOCTEUR ÈS SCIENCES

PAR

Shi-Yan Caroline FOO

acceptée sur proposition du jury:

Prof. M. Blokesch, présidente du jury Prof. S. Cole, directeur de thèse Prof. P. Sander, rapporteur Prof. M. Jackson, rapporteuse Prof. K.-H. Altmann, rapporteur

Suisse 2018

Summary

Tuberculosis (TB), whose etiological agent is Mycobacterium tuberculosis (M. tuberculosis), has plagued humanity since antiquity. Even with chemotherapy available today, TB is the leading cause of death due to an infectious disease. Modern day factors, such as the HIV epidemic and the emergence of drug-resistant TB strains, have redefined the complexities and challenges of tackling the TB pandemic. Current treatments for TB are lengthy, and poor adherence to such prolonged treatment further exacerbates the issue of drug resistance. Moreover, therapies for drug-resistant TB have low cure rates. There is therefore a pressing need for improved therapies that are short and efficacious against all TB strains, which can be achieved through new antimicrobials that are more potent and have novel mechanisms of action, in addition to being affordable, orally bioavailable, and without drug-drug interactions. Such new anti-TB drugs need to be discovered and developed through a long, risky, and costly process, in which attrition rates are high. While there are promising compounds currently being developed, including the benzothiazinones (BTZs), it is necessary to further populate and enhance the Global TB drug pipeline to ensure the availability of new drugs. This thesis aims to address this need through the discovery work of two new, highly promising families, the AX and PB compounds, and of BTZs. The piperazine-based AX analogs are easily synthesised and demonstrate potent activity against M. tuberculosis in vitro and in vivo. Their target, identified in this work, is the QcrB subunit of the cytochrome bc1-aa3 complex, a terminal oxidase of the mycobacterial respiratory chain. Notably, AX compounds are bactericidal in the absence of the alternate terminal oxidase, cytochrome bd. As this family interacts differently in the same binding site of QcrB as Q203, a drug candidate in clinical trials, AX compounds could potentially serve as a backup series for QcrB inhibitors. The PB family, derived from the natural product lapachol, is also easily synthesised and shows substantially improved activity against M. tuberculosis in vitro compared to the parent compound. PB analogs also demonstrated activity against the non-replicating bacillus and in infected macrophages. The mechanism of action of PB compounds relies on the F420 cofactor, although not on the F420-dependent nitroreductase Ddn, therefore this family has a novel mechanism of action which is highly specific to M. tuberculosis. To support the clinical development of PBTZ169, the mechanism of resistance to BTZs was further elucidated in this thesis. Five mutations at cysteine 387 of the target enzyme, DprE1, were identified as mediating resistance to BTZs, which would serve as resistance markers for clinical screening. The impact of these mutations on M. tuberculosis and on DprE1 was further characterised, revealing a fitness cost imposed on the bacillus intracellularly and reduced catalytic efficiency of the enzyme. This thesis additionally contributed to the characterisation of a PBTZ169 backup series and the design of an optimal regimen with other TB drugs. The compounds presented herein merit further optimisation so their full antibiotic potential may be realised for TB, and possibly for other mycobacterial diseases as well. Altogether, this thesis has contributed to the fuelling of drug discovery against M. tuberculosis, and a step towards more medicines for TB.

Keywords: tuberculosis, anti-TB drugs, drug discovery and development, drug resistance, QcrB inhibitors, natural product lapachol, F420 cofactor, benzothiazinone, PBTZ169

Résumé

La tuberculose (TB), dont l'agent étiologique est Mycobacterium tuberculosis (M. tb), affecte l'humanité depuis l'antiquité. Malgré la disponibilité d’une antibiothérapie, elle est la principale cause de décès dû à une maladie infectieuse. Les facteurs modernes, tels que l'épidémie du VIH et l’apparition de souches multi-résistantes, ont redéfini les défis de la lutte contre la pandémie causée par le bacille de Koch. Les traitements actuels sont longs, et le manque d'observance entraîne la résistance aux antibiotiques. De plus, les traitements contre la TB résistante ont de faibles taux de guérison. Il y a donc un besoin urgent d’améliorer les médicaments offerts afin de diminuer le temps de traitement et d’augmenter l’efficacité de ceux-ci contre l’ensemble des souches de M. tb. Par conséquent, la découverte de nouveaux antimycobactériens abordables, ne présentant aucune interaction médicamenteuse, ayant de nouveaux mode d’actions et biodisponibles par voie orale est l’un des nerfs de la guerre afin de contrer la TB dans le monde. Ces nouveaux médicaments sont mis au point dans le cadre d'un processus long et coûteux dans lequel les taux d'attrition sont élevés. Bien que des composés prometteurs soient en cours de développement, dont les benzothiazinones (BTZ), il est nécessaire d’alimenter le pipeline de nouveaux médicaments contre la TB. Cette thèse répond à cette problématique par la découverte de deux nouvelles familles très prometteuses, les composés AX et PB, ainsi que les BTZ. Les analogues AX sont faciles à synthétiser et démontrent une bonne activité contre M. tb in vitro et in vivo. La cible, identifiée dans ce travail, est la sous-unité QcrB du complexe cytochrome bc1-aa3, une oxydase terminale de la chaîne respiratoire mycobactérienne. Les composés AX sont bactéricides en l'absence de l'oxydase terminale alternative, le cytochrome bd. Puisque cette famille interagit différemment au même site de liaison que le Q203, les composés AX pourraient servir de série alternative pour les inhibiteurs de QcrB. La famille PB est dérivée du lapachol, un produit naturel. Les composés sont facilement synthétisables et présentent une amélioration substantielle de l'activité contre M. tb par rapport au lapachol. Les analogues PB ont démontré une activité contre le bacille non-réplicatif ainsi que dans les macrophages infectés. Le mécanisme d'action repose sur le cofacteur F420, mais est indépendant de la nitroréductase Ddn, ce qui suggère que cette famille possède un nouveau mécanisme d'action très spécifique à M. tb. Pour soutenir le développement clinique du PBTZ169, le mécanisme de résistance aux BTZ a été caractérisé. Cinq mutations de la cystéine 387 de l'enzyme cible, DprE1, ont été identifiées et impliquées dans cette résistance. La présence de ces mutations diminue la virulence ex vivo, ainsi qu’elle altère l’efficacité catalytique de l'enzyme Dpre1. Ces mutations pourraient servir de marqueurs de résistance pour le dépistage clinique. Cette thèse a également contribué à la caractérisation d'une série alternative du PBTZ169 et à la conception d'’une combinaison optimale avec d'autres antituberculeux. Les composés présentés ici méritent d’être améliorés afin que leur activité soit optimale contre M. tb et possiblement contre d’autres mycobactéries. Dans l’ensemble, cette thèse a contribué à la découverte de nouveaux antituberculeux afin d’élargir le spectre des outils permettant de lutter contre la TB.

Mots-clés : tuberculose, médicaments antituberculeux, découverte et développement d’antibiotiques, pharmacorésistance, inhibiteur de QcrB, produit naturel lapachol, cofacteur F420, benzothiazinone, PBTZ169

Table of Contents

Abbreviations ...... 3

Chapter 1: Tuberculosis – past and present ...... 5

Chapter 2: Arylvinylpiperazine amides, a new class of potent inhibitors targeting QcrB of

Mycobacterium tuberculosis ...... 63

Chapter 3: Discovery of new plant-derived inhibitors with potent F420-dependent activity against Mycobacterium tuberculosis ...... 129

Chapter 4: Characterization of DprE1-mediated Benzothiazinone Resistance in

Mycobacterium tuberculosis ...... 159

Chapter 5: Conclusions and Perspectives ...... 177

Appendix Chapter A1: Structural studies of Mycobacterium tuberculosis DprE1 interacting with its inhibitors ...... 189

Appendix Chapter A2: Structure-based drug design and characterization of sulfonylpiperazine benzothiazinone inhibitors of DprE1 from Mycobacterium tuberculosis ...... 199

Appendix Chapter A3: An optimized background regimen for treatment of active tuberculosis with the next-generation benzothiazinone Macozinone (PBTZ169) ...... 237

Acknowledgements ...... 271

Curriculum Vitae ...... 273

Abbreviations TB tuberculosis ABC transporters ATP-binding cassette transporters ADME adsorption, distribution, metabolism, and excretion BCG vaccine Bacilli Calmette-Guérin vaccine BDQ bedaquiline, Sirturo BTZs benzothiazinones CFM / CLO clofazimine Ddn deazaflavin-dependent nitroreductase DOT directly-observed therapy DprE1 decaprenylphosphoryl-beta-D-ribose 2’ epimerase DR-TB drug-resistant TB ESX-1 6-kDa early secreted antigen or ESAT-6 secretion system 1 ETC electron transport chain FDA U.S. Food and Drug Administration Fgd1 F420-dependent glucose-6-phosphate dehydrogenase GLP good laboratory practice GMP good manufacturing practices GSK GlaxoSmithKline IFN-Ȗ interferon-Ȗ IGRAs IFN-ȖUHOHDVHDVVD\V INH isoniazid LAM lipoarabinomannan LM lipomannan LPZ(S) lansoprazole (sulphide) LTBI latent TB infection M. bovis Mycobacterium bovis M. tuberculosis Mycobacterium tuberculosis MABSC Mycobacterium abscessus complex MBC minimal bactericidal concentration MCZ Macozinone, PBTZ169 MDR-TB multi-drug resistant TB MIC minimal inhibitory concentration MM4TB More Medicine for Tuberculosis consortium MTBC M. tuberculosis complex NCEs new chemical entities NDH-2 type 2 NADH-quinone oxidoreductase NO nitric oxide NTM non-tuberculous mycobacteria PA-824 Pretomanid PAS para-aminosalicyclic acid PBTZ piperazine-based BTZ PDIM phthiocerol dimycocerosate PK pharmacokinetics PMF proton motive force PPD purified protein derivative PZA pyrazinamide RD-1 region of difference 1

3 RIF rifampicin RND resistance, nodulation and division RNS reactive nitrogen species ROS reactive oxygen species RR-TB rifampicin-resistant TB SAR structure-activity relationship SS18b streptomycin-starved 18b strain TMM or TDM trehalose monomycolate or trehalose dimycolate TST tuberculin skin test WGS whole-genome sequencing WHO World Health Organisation XDR-TB extensively drug-resistant TB

4 Chapter 1: Tuberculosis – past and present

Chapter 1 - Table of Contents

1.1 A Brief History of Tuberculosis ...... 10

1.1.1 TB – an ancient disease ...... 10

1.1.2 Treatment of TB in the pre-antibiotic era ...... 10

1.1.3 New approaches to TB management ...... 11

1.2 Epidemiology ...... 15

1.3 M. tuberculosis, the bug behind the disease ...... 16

1.3.1 Mycobacteria classification ...... 16

1.3.2 Pathogenesis ...... 17

1.3.3 Characteristics of M. tuberculosis ...... 19

1.4 Current approaches to tackling TB ...... 21

1.4.1 Vaccination ...... 21

1.4.2 Diagnosis of TB infection and disease ...... 22

1.4.3 Treatment of TB ...... 24

1.5 The search for new TB drugs ...... 25

1.5.1 The need for new drugs and regimens ...... 25

1.5.2 Strategies for improved TB therapy through new anti-TB drugs ...... 28

1.5.3 TB drug discovery and development ...... 29

1.5.4 Global TB drug pipeline ...... 34

1.6 Thesis rationale and outline ...... 41

References ...... 43

Tuberculosis (TB) is one of the deadliest diseases from antiquity. It is one of the top ten causes of death worldwide today, and remains the leading cause of death due to an infectious disease (1). In 2016 alone, TB claimed 1.7 million lives, contributing to three percent of deaths globally (2). On a broader time scale, the death toll due to TB over the past two centuries is estimated at one billion (3), hence its moniker ‘Captain of all these Men of Death’ (4).

TB, as we know it today, is a contemporary term for the disease. Historically, it has been referred to as consumption, the White Plague, or phthsis by Hipprocrates. Although a disease which primarily affects the lungs, TB also manifests itself in other organs such as the spine

(Pott’s disease), the lymph nodes (scrofula), and the skin (lupus vulgaris).

For all its different labels and forms, the disease caused by Mycobacterium tuberculosis (M. tuberculosis) that has afflicted mankind since thousands of years is the one and the same which persists as a global health problem today. The juxtaposition of an age-old disease against a modern backdrop highlights similarities and contrasts between past and present contexts. While overcrowding and impoverished living conditions (which still unfortunately exist in many areas of the world today) continue to contribute to the spread and mortality of TB, new factors, including the HIV epidemic and drug resistance, add modern complexities to an old problem.

Up until now, the history of TB and humankind has been very much intertwined and will remain so in the foreseeable future. As we continue to seek to tackle TB globally, building upon the work of a great many before us, it is with the aspirations of closing this chapter of history definitively and to put to rest the ‘Captain of all these Men of Death’.

9 1.1 A Brief History of Tuberculosis

1.1.1 TB – an ancient disease

Historical evidence for the existence of TB in humans has been derived from prehistoric art, archaeological artefacts, and literary writings (5). The detection of M. tuberculosis-specific

DNA and lipids in mummy specimens, including those from ancient Egypt (dated 600 BC) (6) and pre-Columbian Peru (dated 1000 AC) (7) provide concrete proof for TB affecting humans in the past. Thus far, the evidence gathered traces TB in humans to more than 8000 years ago with specimens from Atlit Yam in east Mediterranean (5, 8), and paints a picture of a disease with a wide geographical reach with TB epidemics having occurred in ancient Egypt, Greece, the Americas, and Europe (9).

1.1.2 Treatment of TB in the pre-antibiotic era

Treatments for the disease varied greatly throughout different periods of history (9). Ancient

Greek physicians prescribed milk, mild exercise, and warm climates for consumption, while it was believed that the touch of royalty would cure scrofula in Europe during the middle ages.

Common practices during the TB epidemic in Europe in the 17th and 19th centuries included body-cleansing techniques such as sweating, bleeding, and vomiting. Horseback riding and sea voyages were also commonly prescribed to TB patients. These remedies would remain largely ineffective as the White Plague ravaged on.

Sanatoria were associated with TB treatment in the late 19th and early 20th century. Generally considered as the first sanatorium dedicated to TB treatment, Heilanstalt was established in

1859 in the German Silesian Mountains by a German physician, Herman Brehmer (10). Many other TB sanatoria followed suit, including those of Davos (Switzerland) and Seranac Lake

(North America), with Dr. Brehmer’s regimen of bed rest, mountainous outdoor air, exercise

10 and a good diet often being implemented. For the TB patients, there was some success of sanatorium treatment, although its efficacy largely depended on the extent of the disease at the point of hospitalisation (9).

Another procedure performed that was not uncommon during this period was the surgical induction of lung collapse in patients with unilateral TB, on the presumption that limiting lung motions and promoting healing of scarred lung tissue could cure the patient (11). This involved either the injection of air into the pleural cavity (pneumothorax) or removing a portion of the ribs (thoracoplasty). Although this procedure was reported to benefit numerous TB patients, undesirable consequences included embolism, bleeding, and permanent deformity.

While these treatments may have alleviated and treated TB in many cases, and while TB sanatoria contributed to containing the spread of TB from a public health viewpoint, all these practices have since been relegated to history as scientific advancements in the TB field ushered in the antibiotic era.

1.1.3 New approaches to TB management

The defining moment in our understanding of TB arrived with the discovery of its etiological agent, presented by the German scientist Robert Koch in 1882 (12). Previously thought to be hereditary, TB was proven based on Koch’s postulates to be in fact an infectious disease caused by the tubercle bacillus, M. tuberculosis. The immense impact of Koch’s work on TB would later earn him the Nobel Prize in Medicine in 1905.

Koch also created the first diagnostic tool for TB in 1890. Tuberculin, a liquid comprising sterilised, filtered M. tuberculosis culture, was initially aimed at treating TB (13). Although it would prove ineffective for its original purposes, it was further developed into a skin test by an

11 Austrian physician, Clemens Freiherr Baron von Pirquet, to detect prior and ongoing TB infection regardless of disease manifestation (14). The administration of the skin test was further modified by Charles Mantoux such that tuberculin was injected into the skin. The principle active component of Koch’s tuberculin termed as purified protein derivative (PPD) was identified and led to the increased purity and reliability of tuberculin skin tests (15). The batch of PPD produced by Florence Seibert in 1940 still remains as the international reference for the manufacturing of PPD (16).

Eighteen years after Koch’s presentation of his discovery of M. tuberculosis, Albert Calmette and Camille Guérin began work on developing a TB vaccine. After passaging a culture of bovine tubercle bacilli Mycobacterium bovis (M. bovis) 269 times, they obtained a live, attenuated vaccine strain found to be safe in non-human primates (9, 17). The vaccine was named Bacilli Calmette-Guérin (BCG) and was first administered in 1921 in a newborn at high risk of death due to TB, successfully protecting the infant from the disease. Apart from being implicated in a catastrophe in 1930 in Lübeck, Germany, in which a mix-up of strains resulted in children being mistakenly vaccinated with live virulent strains and dying from TB, BCG has demonstrated from widespread vaccination programs to be a safe vaccine, with more than 3 billion doses having been administered worldwide to date (18).

12 Figure 1: A timeline of TB drug discovery and regimen development. Adapted from (19).

The emergence of the TB drug discovery field (Fig. 1) originated on the basis of ‘magic bullets’ designed to specifically target microorganisms without harming the human host. This idea of molecules with selective activity was conceptualised by Paul Ehrlich, and successfully demonstrated when his arsenic-derived drug Salvarsan was used to treat syphilis by targeting

Treponema pallidum in 1910 (20). A breakthrough in anti-TB chemotherapy occurred about 30 years later with the discovery of streptomycin, an antibiotic isolated from a soil actinomycete

Streptomyces griseus (21). Its potent activity was first demonstrated in a guinea pig model of

TB (22), and further validated in the first ever randomised controlled trial conducted in medical history, in which streptomycin was clearly more efficacious than bed rest (23). Resistance to streptomycin emerged within a year of its use (24), which led to the introduction of combination therapy for TB. The first combination of streptomycin and para-aminosalicyclic acid (PAS),

13 another TB drug discovered around the same time as streptomycin, demonstrated that resistance emerged less frequently when both agents were used (25, 26).

More anti-TB drugs were discovered in the 1950s and 1960s. A significant point was the simultaneous discovery of isoniazid (INH) by three pharmaceutical companies, Hoffman-La

Roche, Squibb and Bayer with potent anti-TB activity in vivo (27). This led to the first TB regimen implemented in 1955 consisting of a triple therapy of streptomycin, PAS and INH for

18 to 24 months (28, 29), which would remain as the standard treatment for all forms of TB for the next 15 years (30).

With the discovery of new anti-TB drugs, treatment toxicity was reduced and treatment duration was shortened (Fig. 1). While INH remained the cornerstone for building regimens, major advancements included the replacement of PAS with ethambutol, which reduced toxicity and shortened treatment duration to 18 months, as well as the inclusion of rifampicin (RIF) and pyrazinamide (PZA) which further decreased therapy length to 6 months with reduced relapse rates (31). This 6-month regimen of INH, RIF, PZA, and ethambutol has remained as the front- line TB treatment until today, 38 years after its implementation in the 1980s.

The period between the 1940s and 1960s was a productive era for TB drug discovery, with many significant contributions made by pharmaceutical companies and academics. As TB incidences declined, the disease became less prominent and efforts in TB drug discovery came to a halt. Unfortunately, the late 1980s and early 1990s saw a resurgence in TB cases worldwide driven by the onset of the HIV epidemic and the emergence and transmission of drug-resistant

TB strains (32–34), threatening the efficacy of the front-line regimen. This led to the declaration of TB as a global public health emergency by the World Health Organisation (WHO) in 1993.

14 1.2 Epidemiology (2)

A quarter of the world population has latent TB, defined as an infection with M. tuberculosis without existing clinical disease manifestation. 5 to 15% are predicted to develop active TB within their lifetime, with certain populations at higher risk such as HIV, diabetic, and other immunosuppressed patients. This presents a large potential reservoir of at least 285 million new

TB cases.

In 2016, 10.4 million people fell ill with active TB. 18% of these cases were attributed to undernourishment, 10% to HIV-TB co-infection and about 8% to diabetes. Seven countries accounted for 64% of TB incidence, namely India with the largest incidence, followed by

Indonesia, China, Philippines, Pakistan, Nigeria, and South Africa.

Even with the availability of chemotherapy, TB continues to claim many lives globally. An estimated 1.7 million people died from TB in 2016, of which about 22% were attributed to HIV- associated TB. The HIV-TB co-infection burden is evident: TB is the biggest killer of HIV patients, with 40% of them dying of TB. Since TB remains largely a poverty-related disease,

85% of TB deaths are located in the WHO African and South East Asia Regions, areas of lower socioeconomic status.

Another cause for alarm is the emergence and transmission of drug-resistant TB (DR-TB).

Rifampicin-resistant TB (RR-TB) is TB resistant to RIF, while multi-drug resistant TB (MDR-

TB) is defined as resistance to both RIF and INH, cornerstones of the front-line regiment. Both these cases require second-line treatment regimens comprising fluoroquinolones and injectable agents such as amikacin, capreomycin or kanamycin. Extensively drug-resistant TB (XDR-TB) is defined as MDR-TB with additional resistance to any fluoroquinolone and at least one of the

15 above mentioned injectable agents. In 2016, 600 000 new cases of DR-TB were reported amounting to about 6% of total TB incidence. Of these DR-TB cases, 20% were RR-TB, 80% were MDR-TB cases and 6% of MDR-TB cases were classified as XDR-TB. DR-TB has been reported in 123 countries, with India, China and the Russian Federation alone accounting for half of the cases. Treatment success rates for RR-/MDR-TB and XDR-TB were low, with estimates at 54% and 30%, respectively.

Globally, although there is a decreasing trend in TB mortality rates and incidence rates of 3% and 2% per year, these rates need to be significantly improved in order to meet the milestones of the WHO End TB Strategy as part of the Sustainable Development Goals to end the TB epidemic by 2030.

1.3 M. tuberculosis, the bug behind the disease

1.3.1 Mycobacteria classification

The genus mycobacteria belongs to the actinobacteria family and consists of more than 170 species, including a majority of environmental organisms and several important human pathogens (35). Mycobacteria can be classified as slow- and fast-growers, as defined by colony formation within seven days (36, 37). Most mycobacteria causing diseases in humans are slow- growers, including pathogens within the M. tuberculosis complex (MTBC), M. ulcerans, and

M. leprae, which are the causative agents of TB, Buruli ulcer and leprosy, respectively. The

MTBC comprises human-adapted strains (M. tuberculosis, M. africanum) and animal-adapted strains (e.g. M. bovis which infects cattle, M. microti, the vole bacillus, M. caprae infecting goats, and M. pinipedii that causes TB in seals). Non-tuberculous mycobacteria (NTM), another group of mycobacteria which do not cause TB nor leprosy, are opportunistic environmental

16 pathogens and include slow-growers M. avium and fast-growers of the M. abscessus complex

(MABSC).

1.3.2 Pathogenesis

Figure 2: The cycle of M. tuberculosis infection. The transmission of M. tuberculosis occurs through the inhalation of the mycobacteria in aerosols, whereby an infection is established primarily in the lungs. In most cases, M. tuberculosis is contained by the host immune system in granulomas, resulting in asymptomatic, non-transmissible latent TB. However, there is a risk of the bacillus escaping the granuloma and reactivating, which increases drastically in immunocompromised individuals, such as HIV-TB patients. This results in active TB, in which M. tuberculosis spreads within the host and is transmitted between individuals to establish infections in new hosts. Adapted from (38, 39).

M. tuberculosis is transmitted between individuals in aerosols which are formed when coughing or sneezing, with a single bacterium estimated to be sufficient for establishing an infection (40).

When inhaled, M. tuberculosis cells reach the lung alveoli and are phagocytosed primarily by

17 alveolar macrophages, which are at the front-line of the innate immune response.

Intracellularly, pathogens are typically compartmentalised in the phagosome, which undergoes maturation by fusion with a lysosome to form the phagolysosome. Pathogens face a myriad of host defence mechanisms within the phagolysosome, including an acidic environment, reactive oxygen and nitrogen species (ROS and RNS), lysosomal enzymes, and antimicrobial peptides

(41, 42). However, M. tuberculosis has evolved tactics through a variety of virulence factors to overcome these host defences. It is capable of inhibiting phagolysosome formation and acidification, escaping into the cytosol by lysing the phagosomal membrane, and overcoming xenophagy in the cytosol by preventing the formation of autophagosomes and autolysosomes

(43, 44). In addition, it can induce necrosis instead of apoptosis of the infected macrophage such that upon macrophage death, the bacilli can enter the extracellular space and infect other cells (45).

Infected macrophages produce inflammatory cytokines, triggering the host innate immune response. Neturophils, monocytes, and lymphocytes recruited to the infection site phagocytose more bacteria and amplify the immune signal through the secretion of even more cytokines

(46). As dendritic cells phagocytose the bacilli, they mature and migrate to regional draining lymph nodes to present mycobacterial antigens, which then trigger the adaptive immune response (47). Antigen-specific T-cells are primed and differentiate to effector T-cells, which then migrate back to the infection site (39). CD4+ T-cells have a key role in protective immunity against M. tuberculosis by secreting predominantly interferon-Ȗ (IFN- Ȗ), which activates macrophages to kill intracellular M. tuberculosis by further stimulating antimicrobial processes such as the generation of more RNS (48–50).

18 The entire immune response to the infection results in granuloma formation, with a caseous centre containing lipids and cellular debris of necrotic macrophages surrounded by cellular zones of macrophages, fibroblasts and lymphocytes (51). The granuloma structure is a histopathological hallmark of TB, and this aggregation of immune cells is meant to contain bacilli which cannot be eradicated. Within the granuloma core, M. tuberculosis encounter a variety of microenvironments including low oxygen availability, nutrient deprivation and acidic pH (42, 52). These conditions drive the bacilli into a metabolically-altered dormant state (52), which can last for years.

In most TB-infected individuals, this stalemate between host immune system and pathogen is maintained and enables the containment of bacilli without the manifestation of clinical symptoms. This condition is termed as latent TB infection (LTBI) in which M. tuberculosis is non-transmissible. Immunosuppression, particularly as a result of HIV co-infection, greatly increases the risk of reactivation of the bacilli, leading to their replication and escape from the confinement of the granuloma. They spread in the lung tissue (pulmonary TB), and can also spread through the lymphatic system and blood to other organs (extra-pulmonary or disseminated TB). In this state of active TB, the bacilli can be further transmitted to other hosts, completing the infection cycle of M. tuberculosis (Fig. 2).

1.3.3 Characteristics of M. tuberculosis

M. tuberculosis is a rod-shaped, weakly Gram-staining and acid-fast bacterium with a slow growth rate of 24 h in vitro and in vivo, a highly unusual cell wall, and natural resistance to many antibiotics. The complete sequence of the reference strain M. tuberculosis H37Rv revealed a 4.4 Mbp circular chromosome rich in G + C bases (65%) (53). Its genome indicates a self-sufficient and versatile microorganism as enzymes necessary for the biosynthesis of all

19 essential amino acids, vitamins and enzyme co-factors and for glycolysis, the pentose phosphate pathway, tricarboxylic and glyoxylate metabolic pathways are present, with an unusually large repertoire of enzymes involved in lipid metabolism (53). Furthermore, the genome revealed antibiotic resistance entities such as ȕ-lactamase, aminoglycoside acetyl transferases and drug- efflux systems such as the ATP-binding cassette (ABC) transporters (53).

Figure 3: Schematic representing a simplified view of the mycobacterial cell envelope. Adapted from (54).

The cell wall of mycobacteria differs from those of Gram-positive and Gram-negative bacteria in that apart from the inner cell membrane and core peptidoglycan structure, there are additional arabinogalactan and mycolic acid components (55) (Fig. 3). The arabinogalactan layer is covalently attached to the peptidoglycan layer, and is composed of highly-branched arabinose sugars attached to linear chains of galactose sugars, both in the furanose form (56). Long-chain alpha-alkyl-beta-hydroxy C70-C90 fatty acids form the mycolic acid layer, and these molecules

20 can either be covalently bound to arabinogalactan or esterified to sugars, in particular trehalose

(57). These esters of trehalose, trehalose monomycolate or trehalose dimycolate (TMM or

TDM), are located in the outer membrane segment, which intercalates with mycolic acids (58).

The outer membrane segment also contains non-covalently linked glycophospholipids including lipomannan (LM) and lipoarabinomannan (LAM). Many outer membrane lipids such as TDM, LAM and phthiocerol dimycocerosate (PDIM) have important roles in mediating host interactions and pathogenesis (59–61). Altogether, these layers of polysaccharides, lipids and glycolipids create a highly hydrophobic barrier between the external environment and the bacillus, largely contributing to the natural resistance of M. tuberculosis to many antibiotics.

1.4 Current approaches to tackling TB

1.4.1 Vaccination

One important approach to TB eradication is the vaccination of the population against the pathogen. The only licensed vaccine against TB currently available is BCG, an attenuated form of M. bovis. Its loss of virulence is attributed to the deletion of genomic locus region of difference 1 (RD-1), which encodes components of the ESX-1 (6-kDa early secreted antigen or

ESAT-6 secretion system 1) secretion system, including immunodominant T-cell antigens

EsxA (or ESAT-6) and EsxB (or CFP-10 for 10-kDa cultured filtered protein) (62). Although

BCG highly protects infants and children against disseminated TB, its efficacy against pulmonary TB in the adult population varies widely for reasons which remain unclear (63).

This limited efficacy of BCG thus necessitates the development of new, improved vaccines that protect both children and adults against pulmonary TB.

Current vaccine candidates aim at being administered either upon pre-exposure to M. tuberculosis to protect against active TB, or post-exposure for individuals with LTBI to prevent

21 re-activation (64). Three main strategies are undertaken, mainly the development of live mycobacterial vaccines whereby BCG is improved with the addition of appropriate genes or with the removal of virulence genes from M. tuberculosis, subunit and live vector-based vaccines as boosters after priming by BCG, or killed whole bacterial vaccines (65). Much of the focus of these candidates has been to develop a strong cellular-mediated T-helper 1 cell and

IFN-gamma immune response. However, the lack of correlation between such a response and efficacy as demonstrated by a highly promising vaccine candidate MVA85A failing in Phase

IIb clinical trials underscores the potential need for a more comprehensive set of protective immune markers (66).

1.4.2 Diagnosis of TB infection and disease

Clinical symptoms of pulmonary TB disease include persistent coughing, weight loss, night sweats and fever (67). These symptoms are not specific to TB however, and as individuals with

LTBI are asymptomatic, the diagnosis of TB infection and disease requires a complete medical evaluation. This involves obtaining a medical history of the individual, a physical examination, detection of M. tuberculosis infection, performing a chest radiography, and bacteriological examination of clinical specimens (67).

The detection of M. tuberculosis infection is determined by the Mantoux tuberculin skin test

(TST) and IFN-Ȗ release assays (IGRAs) (68). In the case of TST, PPD is injected under the skin and the diameter of an indurated area is measured 48 to 72 hours later. TST is well- established and extensively used, however several instances can affect its accuracy, including

NTM infections or a previous vaccination with BCG (69). IGRAs such as QuantiFERON-TB

Gold and T-SPOT.TB measure the production of IFN-gamma in the blood induced by TB- specific antigens EsxA and EsxB and are highly specific for M. tuberculosis (68, 69). With the

22 validation of a TB infection, further tests are required to distinguish between LTBI and active

TB.

One test performed is with chest radiographs to detect pulmonary TB, which appears as lung abnormalities such as lesions. Another is the bacteriological examination of clinical specimens, which plays an important role in the diagnosis of TB disease. Sputum smear microscopy involves the staining of acid-fast bacilli from the sputum and detection by microscopy. It is a rapid detection method, although its sensitivity is poor, at around 50%, as more than 104 bacilli per ml of sputum is required for reliable detection (67). Acid-fast organisms other than M. tuberculosis can also generate false-positives (69). A more accurate method of detecting M. tuberculosis is nucleic acid amplification (NAA), which amplifies M. tuberculosis-specific

DNA and RNA segments from clinical specimens. An example is the Xpert MTB/RIF assay

(Cepheid, USA), which is recommended by WHO as a rapid molecular test (2). The reference standard for the confirmation of TB disease is the culturing of specimens for the bacillus (2), which can be performed either on solid growth media or in broth culture systems such as

BACTEC or MGIT (69). This method is highly sensitive, however the long doubling time of

M. tuberculosis leads to a slow diagnosis.

With the diagnosis of TB, drug-susceptibility testing should be conducted. Molecular tests for

DR-TB such as Xpert MTB/RIF assay, HAIN GenoType MTBDRplus and HAIN GenoType

MTBDRsl are useful in providing rapid information for guiding decisions in treatment regimens and resistance surveillance. However, culture-based methods remain the reference standard for the testing of resistance to anti-TB drugs (67).

23 1.4.3 Treatment of TB

The front-line regimen which emerged in the 1980s from drug discovery efforts and clinical trials still remains as the current treatment for drug-susceptible TB (DS-TB). It consists of a six-month, directly-observed therapy (DOT) in which INH, RIF, PZA and ethambutol for the initial two months followed by INH and RIF for the next four months are administered under supervision (2). With complete adherence to the treatment, success rates are at least 85%, and costs US$40 per individual (2).

DR-TB treatment regimens are more complex. Forming a core regimen for MDR-TB treatment involves at least five drugs, including PZA, a fluoroquinolone (levofloxacin, moxifloxacin or gatifloxacin), a second-line injectable (amikacin, kanamycin or capreomycin) and two other core second-line agents such as ethionamide/prothionamide and cycloserine (70, 71). Standard

MDR-TB therapy lasts for 20 months with DOT, including an initial phase of 8 months with the injectable (70). Since 2016, a shorter MDR-TB regimen lasting 9 to 11 months has been recommended by WHO in certain cases. This regimen consists of add-on agents PZA, ethambutol, and high-dose INH in addition to moxifloxacin, kanamycin, prothionamide and clofazimine (70). There is currently no standardised regimen for XDR-TB, and like MDR-TB, regimens are built depending on strain susceptibility and patient tolerability. Altogether, the treatment for DR-TB is costly, more than a hundred times that of DS-TB (2), and long treatment durations are often accompanied by severe side effects.

24 1.5 The search for new TB drugs

1.5.1 The need for new drugs and regimens

There is a pressing need for new TB drugs and regimens, mainly due to two phenomena: (i) resistance towards current antibiotics and (ii) a persistent state of tubercle bacilli which contributes to long treatment durations.

(i) Resistance

Drug-resistant TB either originates from a wild-type strain acquiring resistance within the host

(secondary resistance) or a resistant strain being transmitted between hosts (primary resistance)

(72). Acquired resistance in M. tuberculosis is derived through spontaneous mutations in the chromosome rather than through horizontal gene transfer of mobile elements carrying resistance genes as in other bacteria (73). Such spontaneous mutations occur at a rate estimated at 10-6 to 10-8 replications in M. tuberculosis, and those that impart a survival advantage would be selected for in the presence of a single antibiotic, leading to the multiplication of the resistant strain (74). The probability of such an event occurring theoretically is greatly decreased, to less than 10-24, by combination therapy with four effective TB drugs, that is, effectively negligible

(74). However, sub-optimal combination therapy, in cases for instance of inappropriate prescription, poor compliance, and compartmentalisation of bacilli, increases the likelihood of the selection of spontaneous mutations for resistance towards an antibiotic. In this manner, step- wise acquisition of mutations leads to the emergence of MDR- and XDR-TB strains (75).

Although resistant strains with accumulation of mutations in essential genes should have incurred a certain fitness cost (76), several appear to have retained their virulence and transmissibility with low- or no-cost mutations (77–79), as evident from cases of primary resistance in which individuals having never undergone TB treatment are infected with a

25 resistant strain. Compensatory mutations can also contribute towards lessening the fitness cost of the initial mutation, particularly for initial costly mutations (77), as exemplified by the overexpression of alkyl hydroperoxidase AhpC in INH-resistant strains to compensate for the loss of function of catalase-peroxidase KatG (80).

Table 1 summarises the main genes mediating resistance to front-line and second-line anti-TB agents, as well as their corresponding roles in the mechanism of drug action and cellular processes in which they are involved.

Table 1: Genes involved in resistance to conventional front-line and second-line anti-TB agents. Adapted from (75).

Drug Genes Gene function Role in Cellular process inhibited involved mechanism of in drug action resistance Isoniazid katG Catalase-peroxidase Activator Resistance to oxidative stress

inhA Enoyl acyl carrier protein (ACP) Target Cell wall synthesis reductase (mycolic acid) Rifampicin rpoB RNA polymerase beta-subunit Target RNA synthesis Pyrazinamide pncA Nicotinamidase/pyrazinamidase Activator Membrane energetics rpsA Ribosomal protein S1 Likely Target Trans-translation panD Aspartate decarboxylase Likely Target Pantothenate and co-enzyme A synthesis Ethambutol embB Arabinosyl transferase Target Cell wall synthesis (arabinogalactan) Fluoroquinolones gyrA DNA gyrase subunit A Target DNA synthesis gyrB DNA gyrase subunit B Target Amikacin rrs 16S rRNA of 30S ribosomal Target Protein synthesis subunit Kanamycin rrs 16S rRNA of 30S ribosomal Target Protein synthesis subunit Capreomycin rrs 16S rRNA of 30S ribosomal Target subunit Protein synthesis tlyA 2'-O-methyltransferase Target Streptomycin rpsL S12 ribosomal protein Target Protein synthesis rrs 16S rRNA Target Ethionamide ethA Flavin monoxygenase Activator Cell wall synthesis inhA Enoyl acyl carrier protein (ACP) Target (mycolic acid) reductase Cycloserine alr Alanine racemase Target Cell wall synthesis ddl D-alanine:D-alanine ligase Target (peptidoglycan) PAS thyA Thymidylate synthase Activator? Folate metabolism

26 (ii) Persistence

The idea that distinct sub-populations of tubercle bacilli, existing in different metabolic states during TB infection in the host, was put forward by Mitchison (81). The heterogenous nature of these populations may be a stochastic process, or can arise as a response to drugs, the immune response and/or hostile local microenvironments of low oxygen, nutrient deprivation and acidity (82). Bacilli which are either rapidly multiplying, sporadically multiplying, or are in acidic phagolysosomal compartments are killed by conventional anti-TB drugs (83), which inhibit cellular processes involved during active replication (Table 1).

A sub-population, however, displays drug tolerance and are termed as persisters for their phenotypic resistance to chemotherapy without signs of inheritable genetic mutations (84).

Persistence stems from an altered metabolic state in which bacilli are non-replicating or dormant, and as such these bacilli are difficult to eradicate completely as many targets of conventional anti-TB agents are rendered irrelevant in their case (85). As a result, TB therapy requires a longer continuation phase in addition to an initial intensive phase. Persisters therefore contribute to lengthy treatment durations and increasing the probability of resistance.

In summary, there remains much room for improving current TB treatments. Apart from drug- resistance and long treatment durations, drug-drug interactions of RIF with components of HIV therapy pose problems in HIV-TB cases (86). Moreover, DR-TB regimens are extremely costly and highly toxic with limited efficacy, and the use of injectables is cumbersome for drug administration. Therefore, new regimens comprising new drugs should seek to address these issues.

27 1.5.2 Strategies for improved TB therapy through new anti-TB drugs

Treatment durations can be shortened by using combinations of more potent and sterilising drugs against drug-susceptible and drug-resistant TB strains, which are effective against bacilli in different metabolic and replicating states. This requires drugs with new mechanisms of action. For regimens which are well-tolerated and simple to administer, new drugs need to have good safety profiles with reduced toxicities, as well as high oral bioavailability with preferably not more than once a day dosing. Drugs which are easy to synthesise and have few or no drug- drug interactions would result in more affordable and compatible treatments, such as with antiretroviral therapy for HIV for instance.

Table 2: Goals for new TB regimens and corresponding characteristics of new anti-TB drugs

Desirable attributes of new TB regimens Characteristics of new anti-TB drugs More potent and sterilising drugs, Shorter treatment duration New mechanisms of action Well-tolerated Good safety profile, reduced toxicity Simple administration High oral bioavailability, once a day dosing Affordable Easy synthesis Compatible within and without Few or no drug-drug interactions

In particular, new anti-TB drugs can seek to target either M. tuberculosis as (i) antimicrobials, or (ii) be directed at the host as immunotherapy.

(i) Antimicrobials

For new antimicrobials effective against M. tuberculosis, existing drugs for TB can be further optimised, existing antimicrobials can repurposed for TB, or new chemical entities (NCEs) can be explored for their activity against the pathogen and safety in humans through drug discovery and development (87). This will be discussed in further detail in the next sections.

28 (ii) Host-directed therapy

In the case of host-directed therapy, repurposed drugs and NCEs are assessed for their abilities to modulate host immune responses to eradicate M. tuberculosis more rapidly and/or to prevent lung tissue damage as a consequence of inflammatory processes induced by the immune system reacting to the pathogen (88). Although an important point of consideration in the development of such inhibitors is the potential toxicity due to off-target effects (89), host-directed therapy offers an avenue as adjunct therapy for TB, and in particular for DR-TB.

1.5.3 TB drug discovery and development

Figure 4: Flow of TB drug discovery and development. Some key activities performed at each stage are listed in coloured boxes, with blue being common to both target-based and cell-based approaches and green for the cell-based approach. Adapted from (90).

(i) TB drug discovery: hit-identification, hit-to-lead generation, lead optimisation

Two main strategies can be undertaken for TB drug discovery, namely target-based or phenotypic, whole-cell based approaches (Fig. 4). For both these approaches, the main stages involve hit identification, hit-to-lead generation, and lead optimisation.

29 Hit identification

Identification and validation of the target is the first step towards hit identification in the target- based approach. The complete sequencing of M. tuberculosis has enabled the identification of essential genes for infection in vivo (91, 92). Further validation studies on the essentiality of these genes in vivo and databases such as TuberQ can aid in deciding on a relevant, vulnerable, and druggable target for the drug discovery program (93). Subsequently, screening of compounds is conducted against the target to identify hits. Compounds screened include synthetic, small, drug-like molecules and semi-synthetic, larger molecules derived from natural scaffolds (94). The screening phase can be performed in a high-throughput manner based on compound libraries, either biochemically against the purified enzyme or in silico using its 3D structure (95, 96). Molecules which have a validated activity against the target below a threshold are advanced as hits to the next phase.

For the cell-based approach, hit identification begins with the screening of compounds against the whole bacterial cell and assessing activity of the molecules through growth inhibition or killing of the bacteria (97). Screening conditions, such as growth medium composition (98), models of bacilli mimicking various metabolic states, such as non-replicating M. tuberculosis as induced by microaerophilic and anaerobic conditions in the Wayne model (99), or by withdrawal of streptomycin in the streptomycin-starved 18b strain (SS18b) (100, 101), and intracellular screens (102), are important determinants influencing the susceptibility of M. tuberculosis towards the compounds screened.

A third approach of identifying hits involves target-based, whole-cell screens which combines both of the strategies mentioned above (103). The basis of these screens is the increased

30 susceptibility of a bacterial strain under-expressing the target to the target-specific inhibitor.

Such screens are meant to overcome difficulties often encountered in the subsequent phases of drug discovery such as translation of in vitro activity of hits into whole-cell activity for the target-based approach, and target identification for the cell-based approach (94).

Hit-to-lead generation

Identified hits are advanced to the next stage of hit-to-lead generation. At this stage, hits with promising drug-like features are narrowed down and are refined by medicinal chemistry to improve their potency and selectivity to generate a lead or lead series (104), whereby a lead can be defined as ‘a chemical structure or a series of structures which demonstrate activity and selectivity in a pharmacological or biochemically relevant screen’ (105). Notably, hits identified from target-based screens need to demonstrate activity against whole bacterial cells.

Activities at this stage include more intensive structure-activity-relationship (SAR) studies around the core hit structure which can be rationally guided by the 3D structure of the target and inhibitor-target complex if available, assessing cytotoxicity, determining the minimal bactericidal concentration (MBC) and activity against MDR-TB strains, and profiling of physicochemical and in vitro adsorption, distribution, metabolism, and excretion (ADME) properties (Fig. 4) (90, 104). This information guides generating a lead that has an acceptable, balanced trade-off between potency, selectivity, and physicochemical and ADME properties which have the potential for improvement in the next stage.

In parallel to these activities, studies to identify the target and/or the mechanism of action are undertaken for hits identified through whole-cell screens. With its increasing accessibility and decreasing costs, genomics is a commonly-used technique nowadays, with whole genome

31 sequencing (WGS) of spontaneous resistant mutants to identify targets and/or transcriptomics to aid in deciphering the mechanism of action (106). Validation of the target encompasses confirming the essentiality of the target in vivo, and also genetic and biochemical verification of the specific interaction of the inhibitor with the target. Although the inability to identify the target or the mechanism of action may not necessarily hinder the progress of a compound through the discovery pipeline, it is nevertheless an important step towards understanding its efficacy and safety, both critical parameters for drug development afterwards.

Lead optimisation

The aim of this stage is of lead refinement, in that favourable properties of the lead are maintained while deficiencies of the lead structure are improved upon (104). Lead analogues, which are generated guided by medicinal chemistry, are subjected to further evaluation such as cross-resistance studies with other drugs or leads to verify the target and ex vivo activity in M. tuberculosis-infected macrophages (if this was not determined in the initial screen). In vivo pharmacokinetics (PK) and efficacy studies in mouse TB models give preliminary insight into the metabolism and oral bioavailability of the compound in vivo. At this stage, combination studies may be performed to evaluate potential drug interactions of the lead with other TB drugs or candidates.

A decision is made on the lead analogue to be advanced based on the extensive profiling and evaluation studies performed. During the pre-clinical and clinical development stages of the candidate compound where the risk of failure remains high, drug discovery work continues to generate potential backup molecules and follow-up series (104). Further characterisation studies of the candidate compound such as potential mechanisms of resistance provide more comprehensive insight for its clinical use.

32 (ii) TB drug development: pre-clinical and clinical trials

Pre-clinical studies are designed to assess the efficacy and safety of the pre-clinical candidate using animal models. As they provide a convenient, robust and reproducible model of TB, mice are most commonly used, although studies are also conducted in rats, guinea pigs, rabbits and non-human primates (107). Efficacy is determined alone and in combination with other drugs for assessing potential drug interactions within a regimen (107). To determine a safe window of dosage for human clinical trials, toxicology studies are performed in compliance with good laboratory practice (GLP) (107). In addition, standardising compound formulation and production under good manufacturing practices (GMP) is also required for further clinical development (107).

Figure 5: Stages of clinical trials. The tests conducted, cohort type and size, and costs are indicated for Phase I, II, and III of clinical trials for TB. Costs vary depending on several factors, including the study site, and are estimated based on numbers from (108). Adapted from (109).

Clinical trials are designed to assess the efficacy and safety of the drug candidate in humans.

Phase I trials are conducted in a small number of healthy adults to identify a safe dose range

33 and additional pharmacokinetic data can be obtained from participants (107, 110) (Fig. 5).

Phase II trials are conducted on a larger group of about 100 to 200 TB patients and are meant to determine the efficacy of the compound alone and then in combination with other approved

TB drugs, to establish an optimal dosage, and to gather more pharmacokinetics and safety data

(107, 110) (Fig. 5). An early bactericidal activity (EBA) study is carried out at the beginning of the trial whereby participants are treated with the candidate drug alone over 2 to 14 days to obtain preliminary efficacy results, after which the candidate drug is administered as part of a regimen (111, 112). The current gold-standard for the end-point of the Phase II trials is sputum culture conversion after 2 months (110). Phase III trials are much larger, recruiting in the range of 1000 TB patients (107) (Fig. 5). The aim is to assess the efficacy of the new regimen with the candidate drug and its sterilising ability. As such, patients are followed up for TB relapse for up to 2 years, which is the endpoint for Phase III trials (107).

The entire process from discovering to developing a TB drug is lengthy, costly, complex, and highly risky, calling for much collaborative effort and the involvement of pharmaceutical companies, academic institutions, governmental and non-governmental organisation. The creation of large consortia such as the TB Drug Accelerator (113), the Lilly TB Drug Discovery

(114), and the previously EU-funded More Medicines for Tuberculosis (MM4TB) consortia

(115) dedicated to developing new TB drugs has led to more candidates present in the Global

TB drug pipeline.

1.5.4 Global TB drug pipeline

The current global TB drug and regimen pipeline consists of candidates which are existing TB drugs undergoing optimisation, repurposed drugs or NCEs (Fig. 6).

34 Figure 6: The Global TB Drug Development Pipeline (2018). Repurposed compounds or NCEs currently undergoing pre-clinical and development for TB, as published by the Working Group on New TB drugs in March 2018 (87).

(i) Optimising existing TB drugs/Repurposed drugs

Optimising existing TB drugs and repurposing drugs used in other therapeutic areas are strategies to hasten the availability of new TB drugs. Since these drugs have already been approved by the U.S. Food and Drug Administration, they have validated safety profiles and resources can be diverted to evaluating efficacy against TB. They are also starting scaffolds for improving anti-mycobacterial potency and liabilities of the parent drug.

RIF belongs to the rifamycin class of natural products and is a main sterilising component of front-line TB treatment by binding to bacterial RNA polymerase and inhibiting RNA synthesis

(116). Current clinical trials aim to determine if DS-TB treatment durations can be shortened with higher doses of RIF and with substitution of RIF by rifapentine, a derivative with a longer half-life (117, 118).

35 Fluoroquinolones are broad-spectrum antibiotics repurposed as anti-TB drugs, whereby they inhibit bacterial DNA gyrase, preventing DNA synthesis. Since the 1980s when ciprofloxacin was demonstrated to be useful against DR-TB (119), fluoroquinolones have formed a key component of DR-TB treatment. Today, newer generations of fluoroquinolones such as moxifloxacin, levofloxacin, and gatifloxacin with improved activity against TB are preferred over older generations for MDR-TB treatment (70). Ongoing clinical trials involve assessing moxifloxacin as part of new regimens for DS- and DR-TB treatment (117).

Clofazimine, a member of the riminophenazines, is an anti-leprosy drug being repurposed for treating TB. While its exact mechanism of action is not fully understood, it is considered a prodrug which first requires reduction by type 2 NADH-quinone oxidoreductase (NDH-2) of the mycobacterial electron transport chain (ETC) before releasing ROS upon spontaneous re- oxidation by oxygen (120). Clofazimine is recommended by the WHO as a second-line agent for MDR-TB (70). New riminophenazines, such as TBI-166, are intended to overcome issues of compound solubility and skin discoloration caused by clofazimine (121).

Linezolid is an oxazolidinone used in the treatment of Gram-positive infections (122) which has been repurposed for TB. It is active against M. tuberculosis by targeting the 23S RNA of the 50S ribosomal subunit, inhibiting protein synthesis (123). Linezolid is a recommended second-line agent for MDR-TB treatment, and is evaluated as part of new regimens for DS- and

DR-TB although it has been reported to induce significant toxicities (87, 124). Newer oxazolidinones with improved safety profiles such as sutezolid are being currently developed

(87).

36 Carbapanems belong to the beta-lactam class of broad-spectrum antibiotics inhibiting cell wall

peptidoglycan synthesis. Due to beta-lactamase BlaC in M. tuberculosis limiting the efficacy

of beta-lactams (125), meropenen and faropenem are being assessed in combination with

clavulanate, a beta-lactamase inhibitor, as potential treatment options in clinical trials (126, 87).

(ii) New chemical entities (NCEs)

NCEs contain active moieties which have not been previously approved by the FDA, and thus

have the potential to kill M. tuberculosis through new mechanisms of action. All the NCEs

currently undergoing development in the Global TB drug pipeline have been identified through

cell-based approaches in the discovery stage (Table 3).

Table 3: NCEs in the Global TB drug pipeline. Adapted from (106, 87).

New Chemical Entities (NCEs) Class Compound Hit identification approach Target Mechanism of Action Clinical Comments Development Diarylquinolones Bedaquiline Whole-cell screen of ATP synthase Inhibition of energy production Phase 3 MDR-TB FDA approval, quinolone derivatives add-on agent for MDR- TB

Nitroimidazoles Pretomanid Whole-cell screen of Unknown Inhibition of cell wall synthesis Phase 3 metronidazole derivatives (mycolic acid), respiratory posioning

Delamanid Whole-cell screen of Unknown Inhibition of cell wall synthesis Phase 3 add-on agent for MDR-TB inhibitors of mycolic acid (mycolic acid), likely synthesis respiratory posioning

Diethylamines SQ109 Whole-cell screen of MmpL3 Inhibition of cell wall synthesis Phase 2 ethambutol derivatives (mycolic acid)

Benzothiazinones BTZ043 Whole-cell screen DprE1 Inhibition of cell wall synthesis Pre-clinical O

NO2 (arabinogalactan) O S N

N F3C O Macozinone (PBTZ169) Piperazine-derivative of BTZ DprE1 Inhibition of cell wall synthesis Phase 1, 2 (arabinogalactan)

Imidazopyridines Q203 Whole-cell screen in QcrB Perturbation of respiration Phase 1 M.tuberculosis -infected macrophages

37 Bedaquiline (BDQ, Sirturo) belongs to a new chemical class of the diarylquinolones (Table 3).

It is considered a milestone in the context of contemporary TB drug development as it gained

FDA approval in 2012 and became the first new anti-TB drug to be approved in more than 40 years (127). The target of BDQ was identified by WGS of spontaneous resistant mutants isolated in M. tuberculosis, which revealed missense mutations in atpE encoding for the c subunit of ATP synthase (128, 129). BDQ acts by inhibiting ATP synthesis, an essential process for both replicating and non-replicating states of M. tuberculosis (130), and thus is bactericidal against these populations albeit with a delayed onset of killing (131). Not all clinical isolates resistant to BDQ have mutations in atpE, implying that there is another target or that there are other mechanisms of resistance (132). This has been demonstrated as upregulation of MmpL5, a multi-substrate efflux pump of the MmpL proteins, causes resistance to both BDQ and clofazimine (133, 134). Recommended as an add-on agent for MDR-TB treatment, BDQ has a long oral bioavailability, however it has been associated with the potential induction of arrhythmia and an increased risk of death (127).

Delamanid and PA-824 (also known as Pretomanid) (Table 3) are two new nitroimidazoles identified independently in whole-cell screens (135, 136). Both compounds are active against

MDR-TB and against replicating and non-replicating M. tuberculosis. They have similar mechanisms of action, in that their aerobic activity is attributed to the inhibition of mycolic acid synthesis (135, 137). The anaerobic activity of PA-824 relies on its nitro-reduction by F420- deazaflavin-dependent nitroreductase (Ddn) of M. tuberculosis (138), which results in the release of nitric oxide (NO) causing subsequent respiratory poisoning of the pathogen (139).

Mutations in enzymes of the F420 cofactor biosynthesis pathway, or in F420-dependent glucose-6-phosphate dehydrogenase (Fgd1) which generates reduced F420 required for Ddn

38 activity, or in Ddn itself are associated with resistance to these compounds (140, 141). Both are in advanced clinical trials, with Delamanid having been approved for MDR-TB treatment by the European Commission (142).

SQ109 is an ethylenediamine-based analogue (Table 3) originating from a whole-cell screen which aimed at generating more potent diamines than ethambutol (143). A target of SQ109 was identified as MmpL3 (144), an essential protein belonging to the resistance, nodulation and division (RND) family which transports trehalose monomycolate to the cell envelope for mycolic acid synthesis (145). Apart from inhibiting cell wall synthesis, it has been demonstrated that SQ109 also interferes with menaquinone synthesis and dissipates the proton motive force

(PMF) across the mycobacterial membrane, resulting in perturbation of respiration (146). This multi-modal mechanism of action would imply lower resistance frequencies to SQ109, which is currently being developed in Phase II clinical trials (87).

Benzothiazinones (BTZs) (Table 3) are a new class of extremely potent inhibitors against M. tuberculosis, with the original lead BTZ043 identified in a whole-cell screen (147), and improved upon with a more potent, safer piperazine-containing benzothiazinone PBTZ169

(148). These BTZs have low nanomolar activity against mycobacteria in vitro and ex vivo and are active against DS- and DR-TB clinical isolates (147–149). Through genetic, biochemical, transcriptomic and proteomic analyses, the target was identified and validated as decaprenylphosphoryl-beta-D-ribose 2’ epimerase (DprE1) (147). DprE1 is an essential flavoenzyme of M. tuberculosis, and acts in concert with DprE2 to convert decaprenylphosphoryl-D-ribose to decaprenyl-phosphoryl-D-arabinose, the sole precursor for the biosynthesis of the arabinogalactan and lipoarabinomannan components of the mycobacterial cell wall (150). BTZs are prodrugs requiring activation by reduced FAD of DprE1 (151). The

39 nitro group undergoes nitroreduction to form a nitroso species, which covalently binds to cysteine 387 in the active site of DprE1 (152–154). This irreversible interaction results in the inhibition of cell wall arabinan synthesis, causing lysis of the mycobacterial cell (147). Synergy has been demonstrated for PBTZ169 with BDQ, which is promising for a novel TB regimen

(148, 155). PBTZ169 is undergoing Phase I and II clinical trials, while BTZ043 is in pre-clinical trials as a backup compound (87).

Q203, the most advanced candidate of the imidazopyridine amides class (Table 3), was identified in a whole-cell screen in M. tuberculosis-infected macrophages (156). It has an MIC in the nanomolar range, and is active against MDR-TB clinical isolates (156). Resistance to

Q203 is mediated by mutations in QcrB (156), the b subunit of cytochrome bc1 complex, one of two terminal oxidases in mycobacteria (157). Inhibition of QcrB by Q203 results in rapid depletion of intracellular ATP levels (156) and enhanced killing of M. tuberculosis in mice with the genetic deletion of the alternate terminal oxidase cytochrome bd (158). This emphasises the extreme vulnerability of the bacillus when both terminal respiratory oxidases are simultaneously targeted. With its good oral bioavailability in mice and long half-life, Q203 is undergoing Phase I clinical trials (87).

Although promising compounds with new mechanisms of action are currently present in the

Global TB drug pipeline, there remains nevertheless a need to enhance the pipeline due to the high attrition rates in pre-clinical development and clinical trials that is the very nature of drug discovery and development.

40 1.6 Thesis rationale and outline

The broader aim of this thesis is to address the ongoing global health issue by contributing to the search for new, efficacious, and safe TB antimicrobials. In particular, the scope of this thesis is to enhance the Global TB drug pipeline through the discovery work on several new compounds, including BTZs.

Chapter 2 focuses on hit/lead optimisation activities of new piperazine-based AX compounds, of which the lead molecule, AX-35, was previously identified in a large phenotypic screening campaign conducted by GlaxoSmithKline (GSK) in 2013. Due to its micromolar potency against M. tuberculosis and structural simplicity, AX-35 was deemed as an attractive starting point for lead analogue generation. In this chapter, the characterisation of AX-35 and its four most potent analogues are described, as well as target identification and validation studies undertaken to elucidate their mechanism of action.

The ongoing discovery work of new derivatives of lapachol, the PB compounds, is detailed in

Chapter 3. Lapachol is a natural product originating from trees of the Bignoniaceae family, and its availability and simple structure provides an attractive starting scaffold for analogue generation. PB compounds demonstrate potent activity against M. tuberculosis in vitro in the micromolar range, thus forming the basis of further characterisation work of these series, including some insight into their mechanism of action.

Continued discovery work on the BTZs, BTZ043 and PBTZ169 (Macozinone, MCZ), is intended to support their development as they undergo pre-clinical and phase I/II clinical trials respectively. One aspect is in the understanding of the mechanism of resistance, which can enable the identification of a diagnostic marker for drug-susceptibility and can be screened for

41 over the course of clinical trials to monitor acquired resistance. Through a comprehensive biological and biochemical characterisation of BTZ-resistant mutants, Chapter 4 sheds more light on the underlying mechanisms of BTZ resistance and the impact of these mutations at cysteine 387 of DprE1 on M. tuberculosis fitness, DprE1 enzymatic activity, and the binding of covalent and non-covalent DprE1 inhibitors.

To address potential resistance issues with MCZ and as part of the MCZ back-up program, sulfone-containing derivatives were envisioned by structure-based drug design then generated, with the aim of increasing compound activity against wild-type and BTZ-resistant M. tuberculosis and improving physicochemical properties of MCZ. Chapters A1 and A2 of the annex describes the rationale of generating the sulfone-containing MCZ series and the comprehensive characterisation of these derivatives.

Since TB treatment requires combination therapy and as new TB regimens are needed to improve upon existing ones, Chapter A3 of the annex seeks to explore the potential of MCZ as a part of novel TB regimens by a thorough evaluation of its interactions with other conventional and new TB drugs through in vitro and in vivo combination studies.

Lastly, Chapter 5 concludes the thesis with a summary of the main contributions, as well as some perspectives on the future directions of the work presented here.

42 References

1. 2017. The top 10 causes of death. WHO.

2. World Health Organization. 2017. Global Tuberculosis Report 2017. S.l.

3. Koehler CSW. 2002. Consumption, the great killer. Mod Drug Discov 5.

4. Bunyan J. 1680. The Life And Death Of Mr Badman (The Twin Book To The Pilgrim’s

Progress).

5. Donoghue HD. 2016. Paleomicrobiology of Human Tuberculosis, p. 113–130. In

Drancourt, M, Raoult, D (eds.), Paleomicrobiology of Humans. American Society of

Microbiology.

6. Donoghue HD, Lee OY-C, Minnikin DE, Besra GS, Taylor JH, Spigelman M. 2010.

Tuberculosis in Dr Granville’s mummy: a molecular re-examination of the earliest known

Egyptian mummy to be scientifically examined and given a medical diagnosis. Proc R Soc

B Biol Sci 277:51–56.

7. Salo WL, Aufderheide AC, Buikstra J, Holcomb TA. 1994. Identification of

Mycobacterium tuberculosis DNA in a pre-Columbian Peruvian mummy. Proc Natl Acad

Sci 91:2091–2094.

8. Hershkovitz I, Donoghue HD, Minnikin DE, May H, Lee OY-C, Feldman M, Galili E,

Spigelman M, Rothschild BM, Bar-Gal GK. 2015. Tuberculosis origin: The Neolithic

scenario. Tuberculosis 95:S122–S126.

9. Daniel TM. 1999. Captain of Death: The Story of Tuberculosis.

43 10. Daniel TM. 2011. Hermann Brehmer and the origins of tuberculosis sanatoria [Founders of

our knowledge]. Int J Tuberc Lung Dis 15:161–162.

11. The Surgery for Pulmonary Tuberculosis | American Review of Respiratory Disease.

12. Koch R. 2010. Die aetiologie der tuberkulose.

13. Koch R. 1890. A Further Communication on a Remedy for Tuberculosis. Br Med J 2:1193–

1199.

14. Pirquet CV. 1909. Frequency of Tuberculosis in Childhood. J Am Med Assoc LII:675–678.

15. Long ER, Seibert FB, Aronson JD. 1934. A standardised tuberculin (purified protein

derivative) for uniformity in diagnosis and epidemiology. Tubercle 16:304–322.

16. Seibert FB, Glenn JT. 1941. Tuberculin Purified Protein Derivative. Preparation and

Analyses of a Large Quantity for Standard. Am Rev Tuberc Pulm Dis 44:9–25.

17. Bloom BR, Fine PEM. 1994. The BCG Experience: Implications for Future Vaccines

against TuberculosisTuberculosis: Pathogenesis, Protection, and Control. American

Society of Microbiology, Washington, DC 200005.

18. SAGE Working Group on BCG Vaccines, WHO. 2017. Report on BCG vaccine use for

protection against mycobacterial infections including tuberculosis, leprosy, and other

nontuberculous mycobacteria (NTM) infections.

19. Ma Z, Lienhardt C, McIlleron H, Nunn AJ, Wang X. 2010. Global tuberculosis drug

development pipeline: the need and the reality. The Lancet 375:2100–2109.

20. Thorburn AL. 1983. Paul Ehrlich: pioneer of chemotherapy and cure by arsenic (1854-

1915). Sex Transm Infect 59:404–405.

44 21. Schatz A, Bugle E, Waksman SA. 1944. Streptomycin, a Substance Exhibiting Antibiotic

Proc Soc Exp Biol Med .†כ.Activity Against Gram-Positive and Gram-Negative Bacteria

55:66–69.

22. Feldman WH, Karlson AG, Hinshaw HC. 1947. Streptomycin in Experimental

Tuberculosis. The Effects in Guinea Pigs following Infection by Intravenous Inoculation.

Am Rev Tuberc Pulm Dis 56:346–59.

23. 1948. Streptomycin Treatment of Pulmonary Tuberculosis. Br Med J 2:769–782.

24. Hinshaw C, Feldman WH, Pfuetze KH. 1946. Treatment of Tuberculosis with

Streptomycin: A Summary of Observations on One Hundred Cases. J Am Med Assoc

132:778–782.

25. Lehmann J. 1946. Para-aminosalicylic acid in the treatment of tuberculosis. The Lancet

247:15–16.

26. 1950. Treatment of Pulmonary Tuberculosis with Streptomycin and Para-Amino-Salicylic

Acid. Br Med J 2:1073–1085.

27. Bernstein J, Lott WA, Steinberg BA, Yale HL. 1952. Chemotherapy of Experimental

Tuberculosis. V. Isonicotinic Acid Hydrazide (Nydrazid) and Related Compounds. Am Rev

Tuberc Pulm Dis 65:357–64.

28. Group BMJP. 1955. Various Combinations of Isoniazid with Streptomycin or with P.A.S.

in the Treatment of Pulmonary Tuberculosis: Seventh Report to the Medical Research

Council. Br Med J 1:435–445.

29. Iseman MD. 2002. Tuberculosis therapy: past, present and future. Eur Respir J 20:87S–94s.

45 30. Murray JF, Schraufnagel DE, Hopewell PC. 2015. Treatment of Tuberculosis. A Historical

Perspective. Ann Am Thorac Soc 12:1749–1759.

31. Fox W, Ellard GA, Mitchison DA. 1999. Studies on the treatment of tuberculosis

undertaken by the British Medical Research Council Tuberculosis Units, 1946 to 1986, with

relevant subsequent publications. Text.

32. Giovanni D, Danzi MC, Checchi GD, Pizzighella S, Solbiati M, Cruciani M, Luzzati R,

Malena M, Mazzi R, Concia E, Bassetti D. 1989. Nosocomial Epidemic of Active

Tuberculosis Among HIV-infected Patients. The Lancet 334:1502–1504.

33. Daley CL, Small PM, Schecter GF, Schoolnik GK, McAdam RA, Jacobs WRJ, Hopewell

PC. 2010. An Outbreak of Tuberculosis with Accelerated Progression among Persons

Infected with the Human Immunodeficiency Virus.

http://dx.doi.org/101056/NEJM199201233260404. research-article.

34. Frieden TR, Sterling T, Pablos-Mendez A, Kilburn JO, Cauthen GM, Dooley SW. 2010.

The Emergence of Drug-Resistant Tuberculosis in New York City.

http://dx.doi.org/101056/NEJM199302253280801. research-article.

35. Gagneux S. 2018. Ecology and evolution of Mycobacterium tuberculosis.Nat Rev

Microbiol 16:202–213.

36. Springer B, Stockman L, Teschner K, Roberts GD, Böttger EC. 1996. Two-laboratory

collaborative study on identification of mycobacteria: molecular versus phenotypic

methods. J Clin Microbiol 34:296–303.

37. Rastogi N, Legrand E, Sola C. 2001. The mycobacteria: an introduction to nomenclature

and pathogenesis. Rev Sci Tech-Off Int Epizoot 20:21–54.

46 38. Kaufmann SHE. 2001. How can immunology contribute to the control of tuberculosis? Nat

Rev Immunol 1:20–30.

39. Nunes-Alves C, Booty MG, Carpenter SM, Jayaraman P, Rothchild AC, Behar SM. 2014.

In search of a new paradigm for protective immunity to TB. Nat Rev Microbiol 12:289–

299.

40. Russell DG, Barry CE, Flynn JL. 2010. Tuberculosis: What We Don’t Know Can, and

Does, Hurt Us. Science 328:852–856.

41. Nathan CF, Hibbs Jr JB. 1991. Role of nitric oxide synthesis in macrophage antimicrobial

activity. Curr Opin Immunol 3:65–70.

42. Smith I. 2003. Mycobacterium tuberculosis Pathogenesis and Molecular Determinants of

Virulence. Clin Microbiol Rev 16:463–496.

43. Frehel C, Chastellier C de, Lang T, Rastogi N. 1986. Evidence for inhibition of fusion of

lysosomal and prelysosomal compartments with phagosomes in macrophages infected with

pathogenic Mycobacterium avium. Infect Immun 52:252–262.

44. Crowle AJ, Dahl R, Ross E, May MH. 1991. Evidence that vesicles containing living,

virulent Mycobacterium tuberculosis or Mycobacterium avium in cultured human

macrophages are not acidic. Infect Immun 59:1823–1831.

45. Philips JA, Ernst JD. 2012. Tuberculosis Pathogenesis and Immunity. Annu Rev Pathol

Mech Dis 7:353–384.

46. Crevel R van, Ottenhoff THM, Meer JWM van der. 2002. Innate Immunity to

Mycobacterium tuberculosis. Clin Microbiol Rev 15:294–309.

47 47. Wolf AJ, Desvignes L, Linas B, Banaiee N, Tamura T, Takatsu K, Ernst JD. 2008. Initiation

of the adaptive immune response to Mycobacterium tuberculosis depends on antigen

production in the local lymph node, not the lungs. J Exp Med 205:105–115.

48. Orme IM, Roberts AD, Griffin JP, Abrams JS. 1993. Cytokine secretion by CD4 T

lymphocytes acquired in response to Mycobacterium tuberculosis infection. J Immunol

151:518–525.

49. Flynn JL, Chan J, Triebold KJ, Dalton DK, Stewart TA, Bloom BR. 1993. An essential role

for interferon gamma in resistance to Mycobacterium tuberculosis infection. J Exp Med

178:2249–2254.

50. Chan J, Xing Y, Magliozzo RS, Bloom BR. 1992. Killing of virulent Mycobacterium

tuberculosis by reactive nitrogen intermediates produced by activated murine macrophages.

J Exp Med 175:1111–1122.

51. Russell DG. 2007. Who puts the tubercle in tuberculosis? Nat Rev Microbiol 5:39–47.

52. Timm J, Post FA, Bekker L-G, Walther GB, Wainwright HC, Manganelli R, Chan W-T,

Tsenova L, Gold B, Smith I, Kaplan G, McKinney JD. 2003. Differential expression of

iron-, carbon-, and oxygen-responsive mycobacterial genes in the lungs of chronically

infected mice and tuberculosis patients. Proc Natl Acad Sci 100:14321–14326.

53. Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D, Gordon SV, Eiglmeier K,

Gas S, Iii CEB, Tekaia F, Badcock K, Basham D, Brown D, Chillingworth T, Connor R,

Davies R, Devlin K, Feltwell T, Gentles S, Hamlin N, Holroyd S, Hornsby T, Jagels K,

Krogh A, McLean J, Moule S, Murphy L, Oliver K, Osborne J, Quail MA, Rajandream M-

A, Rogers J, Rutter S, Seeger K, Skelton J, Squares R, Squares S, Sulston JE, Taylor K,

48 Whitehead S, Barrell BG. 1998. Deciphering the biology of Mycobacterium tuberculosis

from the complete genome sequence. Nature 393:537–544.

54. Brown L, Wolf JM, Prados-Rosales R, Casadevall A. 2015. Through the wall: extracellular

vesicles in Gram-positive bacteria, mycobacteria and fungi. Nat Rev Microbiol 13:620–

630.

55. Brennan PJ, Nikaido H. 1995. The Envelope of Mycobacteria. Annu Rev Biochem 64:29–

63.

56. McNeil M, Wallner SJ, Hunter SW, Brennan PJ. 1987. Demonstration that the galactosyl

and arabinosyl residues in the cell-wall arabinogalactan of Mycobacterium leprae and

Myobacterium tuberculosis are furanoid. Carbohydr Res 166:299–308.

57. Cole ST. 2012. Infectious diseases: Transporter targeted in tuberculosis. Nat Chem Biol

8:326–327.

58. Jankute M, Cox JAG, Harrison J, Besra GS. 2015. Assembly of the Mycobacterial Cell

Wall. Annu Rev Microbiol 69:405–423.

59. Briken V, Porcelli SA, Besra GS, Kremer L. 2004. Mycobacterial lipoarabinomannan and

related lipoglycans: from biogenesis to modulation of the immune response. Mol Microbiol

53:391–403.

60. Camacho LR, Ensergueix D, Perez E, Gicquel B, Guilhot C. Identification of a virulence

gene cluster of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis.

Mol Microbiol 34:257–267.

61. Cox JS, Chen B, McNeil M, Jr WRJ. 1999. Complex lipid determines tissue-specific

replication of Mycobacterium tuberculosis in mice. Nature 402:79–83.

49 62. Pym AS, Brodin P, Brosch R, Huerre M, Cole ST. 2002. Loss of RD1 contributed to the

attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and

Mycobacterium microti. Mol Microbiol 46:709–717.

63. Colditz GA, Brewer TF, Berkey CS, Wilson ME, Burdick E, Fineberg HV, Mosteller F.

1994. Efficacy of BCG Vaccine in the Prevention of Tuberculosis: Meta-analysis of the

Published Literature. JAMA 271:698–702.

64. Andersen P, Woodworth JS. 2014. Tuberculosis vaccines – rethinking the current

paradigm. Trends Immunol 35:387–395.

65. Kaufmann SH, Hussey G, Lambert P-H. 2010. New vaccines for tuberculosis. The Lancet

375:2110–2119.

66. Tameris MD, Hatherill M, Landry BS, Scriba TJ, Snowden MA, Lockhart S, Shea JE,

McClain JB, Hussey GD, Hanekom WA, Mahomed H, McShane H. 2013. Safety and

efficacy of MVA85A, a new tuberculosis vaccine, in infants previously vaccinated with

BCG: a randomised, placebo-controlled phase 2b trial. The Lancet 381:1021–1028.

67. Centers for Disease Control and Prevention. Diagnosis of Tuberculosis DiseaseDiagnosis

of Tuberculosis Disease.

68. Centers for Disease Control and Prevention. Testing for Tuberculosis Infection and

DiseaseTesting for Tuberculosis Infection and Disease.

69. Dorman SE. 2010. New Diagnostic Tests for Tuberculosis: Bench, Bedside, and Beyond.

Clin Infect Dis 50:S173–S177.

70. World Health Organization, Global Tuberculosis Programme. 2016. WHO treatment

guidelines for drug-resistant tuberculosis: 2016 update.

50 71. 2015. Companion Handbook to the 2011 Who Guidelines for the Programmatic

Management of Multidrug-resistant Tuberculosis. World Health Organization.

72. Kochi A, Vareldzis B, Styblo K. 1993. Multidrug-resistant tuberculosis and its control. Res

Microbiol 144:104–110.

73. McGrath M, Pittius G van, C N, Helden V, D P, Warren RM, Warner DF. 2014. Mutation

rate and the emergence of drug resistance in Mycobacterium tuberculosis. J Antimicrob

Chemother 69:292–302.

74. Gillespie SH. 2002. Evolution of Drug Resistance in Mycobacterium tuberculosis: Clinical

and Molecular Perspective. Antimicrob Agents Chemother 46:267–274.

75. Zhang Y, Yew WW. 2009. Mechanisms of drug resistance in Mycobacterium tuberculosis

[State of the art series. Drug-resistant tuberculosis. Edited by CY. Chiang. Number 1 in the

series]. Int J Tuberc Lung Dis 13:1320–1330.

76. Spratt BG. 1996. Antibiotic resistance: Counting the cost. Curr Biol 6:1219–1221.

77. Sander P, Springer B, Prammananan T, Sturmfels A, Kappler M, Pletschette M, Böttger

EC. 2002. Fitness Cost of Chromosomal Drug Resistance-Conferring Mutations.

Antimicrob Agents Chemother 46:1204–1211.

78. Davies A., Billington O., Bannister B., Weir WR., McHugh T., Gillespie S. 2000.

Comparison of Fitness of Two Isolates of Mycobacterium tuberculosis, one of Which had

Developed Multi-drug Resistance During the Course of Treatment. J Infect 41:184–187.

79. Gagneux S, Long CD, Small PM, Van T, Schoolnik GK, Bohannan BJM. 2006. The

Competitive Cost of Antibiotic Resistance in Mycobacterium tuberculosis. Science

312:1944–1946.

51 80. Sherman DR, Mdluli K, Hickey MJ, Arain TM, Morris SL, Barry CE, Stover CK. 1996.

Compensatory ahpC Gene Expression in Isoniazid-Resistant Mycobacterium tuberculosis.

Science 272:1641–1643.

81. Mitchison DA. 1979. Basic Mechanisms of Chemotherapy. Chest 76:771–780.

82. Sacchettini JC, Rubin EJ, Freundlich JS. 2008. Drugs versus bugs: in pursuit of the

persistent predator Mycobacterium tuberculosis. Nat Rev Microbiol 6:41–52.

83. McKinney JD. 2000. In vivo veritas: The search for TB drug targets goes live. Nat Med

6:1330–1333.

84. Dhar N, McKinney JD. 2007. Microbial phenotypic heterogeneity and antibiotic tolerance.

Curr Opin Microbiol 10:30–38.

85. Gomez JE, McKinney JD. 2004. M. tuberculosis persistence, latency, and drug tolerance.

Tuberculosis 84:29–44.

86. McIlleron H, Meintjes G, Burman WJ, Maartens G. 2007. Complications of Antiretroviral

Therapy in Patients with Tuberculosis: Drug Interactions, Toxicity, and Immune

Reconstitution Inflammatory Syndrome. J Infect Dis 196:S63–S75.

87. Working Group for New TB Drugs |.

88. Wallis RS, Maeurer M, Mwaba P, Chakaya J, Rustomjee R, Migliori GB, Marais B, Schito

M, Churchyard G, Swaminathan S, Hoelscher M, Zumla A. 2016. Tuberculosis—advances

in development of new drugs, treatment regimens, host-directed therapies, and biomarkers.

Lancet Infect Dis 16:e34–e46.

52 89. Lechartier B, Rybniker J, Zumla A, Cole ST. 2014. Tuberculosis drug discovery in the post-

post-genomic era. EMBO Mol Med 6:158–168.

90. Manjunatha UH, Smith PW. 2015. Perspective: Challenges and opportunities in TB drug

discovery from phenotypic screening. Bioorg Med Chem 23:5087–5097.

91. McKinney JD, Bentrup KH zu, Muñoz-Elías EJ, Miczak A, Chen B, Chan W-T, Swenson

D, Sacchettini JC, Jr WRJ, Russell DG. 2000. Persistence of Mycobacterium tuberculosis

in macrophages and mice requires the glyoxylate shunt enzyme isocitrate lyase. Nature

406:735–738.

92. Sassetti CM, Rubin EJ. 2003. Genetic requirements for mycobacterial survival during

infection. Proc Natl Acad Sci U S A 100:12989–12994.

93. Radusky L, Defelipe LA, Lanzarotti E, Luque J, Barril X, Marti MA, Turjanski AG. 2014.

TuberQ: a Mycobacterium tuberculosis protein druggability database. Database 2014.

94. Zuniga ES, Early J, Parish T. 2015. The future for early-stage tuberculosis drug discovery.

Future Microbiol 10:217–229.

95. Wilsey C, Gurka J, Toth D, Franco J. 2013. A large scale virtual screen of DprE1. Comput

Biol Chem 47:121–125.

96. Pauli I, dos Santos RN, Rostirolla DC, Martinelli LK, Ducati RG, Timmers LFSM, Basso

LA, Santos DS, Guido RVC, Andricopulo AD, Norberto de Souza O. 2013. Discovery of

New Inhibitors of Mycobacterium tuberculosis InhA Enzyme Using Virtual Screening and

a 3D-Pharmacophore-Based Approach. J Chem Inf Model 53:2390–2401.

97. Sala C, Hartkoorn RC. 2011. Tuberculosis drugs: new candidates and how to find more.

Future Microbiol 6:617–633.

53 98. Franzblau SG, DeGroote MA, Cho SH, Andries K, Nuermberger E, Orme IM, Mdluli K,

Angulo-Barturen I, Dick T, Dartois V, Lenaerts AJ. 2012. Comprehensive analysis of

methods used for the evaluation of compounds against Mycobacterium tuberculosis.

Tuberculosis 92:453–488.

99. Wayne LG, Hayes LG. 1996. An in vitro model for sequential study of shiftdown of

Mycobacterium tuberculosis through two stages of nonreplicating persistence. Infect

Immun 64:2062–2069.

100. Deb C, Lee C-M, Dubey VS, Daniel J, Abomoelak B, Sirakova TD, Pawar S, Rogers

L, Kolattukudy PE. 2009. A Novel In Vitro Multiple-Stress Dormancy Model for

Mycobacterium tuberculosis Generates a Lipid-Loaded, Drug-Tolerant, Dormant

Pathogen. PLOS ONE 4:e6077.

101. Sala C, Dhar N, Hartkoorn RC, Zhang M, Ha YH, Schneider P, Cole ST. 2010. Simple

Model for Testing Drugs against Nonreplicating Mycobacterium tuberculosis. Antimicrob

Agents Chemother 54:4150–4158.

102. Christophe T, Jackson M, Jeon HK, Fenistein D, Contreras-Dominguez M, Kim J,

Genovesio A, Carralot J-P, Ewann F, Kim EH, Lee SY, Kang S, Seo MJ, Park EJ,

Škovierová H, Pham H, Riccardi G, Nam JY, Marsollier L, Kempf M, Joly-Guillou M-L,

Oh T, Shin WK, No Z, Nehrbass U, Brosch R, Cole ST, Brodin P. 2009. High Content

Screening Identifies Decaprenyl-3KRVSKRULERVHƍ(SLPHUDVHDVD7DUJHWIRU,QWUDFHOOXODU

Antimycobacterial Inhibitors. PLOS Pathog 5:e1000645.

103. Abrahams GL, Kumar A, Savvi S, Hung AW, Wen S, Abell C, Barry CE, Sherman DR,

Boshoff HIM, Mizrahi V. 2012. Pathway-Selective Sensitization of Mycobacterium

tuberculosis for Target-Based Whole-Cell Screening. Chem Biol 19:844–854.

54 104. Hughes JP, Rees S, Kalindjian SB, Philpott KL. 2011. Principles of early drug

discovery. Br J Pharmacol 162:1239–1249.

105. Bleicher KH, Böhm H-J, Müller K, Alanine AI. 2003. A guide to drug discovery: Hit

and lead generation: beyond high-throughput screening. Nat Rev Drug Discov 2:369–378.

106. Lechartier B, Rybniker J, Zumla A, Cole ST. 2014. Tuberculosis drug discovery in the

post-post-genomic era. EMBO Mol Med n/a-n/a.

107. Global Alliance for TB Drug Development. 2001. Tuberculosis. Scientific blueprint for

tuberculosis drug development. Tuberc Edinb Scotl 81 Suppl 1:1–52.

108. Schluger N, Karunakara U, Lienhardt C, Nyirenda T, Chaisson R. 2007. Building

Clinical Trials Capacity for Tuberculosis Drugs in High-Burden Countries. PLOS Med

4:e302.

109. Dorman S. New Drugs, New Regimens, Clinical Trials.

110. Bark CM, Furin JJ, Johnson JL. 2012. Approaches to clinical trials of new anti-TB

drugs. Clin Investig 2:359–370.

111. Jindani A, Aber VR, Edwards EA, Mitchison DA. 1980. The Early Bactericidal Activity

of Drugs in Patients with Pulmonary Tuberculosis. Am Rev Respir Dis 121:939–949.

112. Donald PR, Diacon AH. 2008. The early bactericidal activity of anti-tuberculosis drugs:

a literature review. Tuberculosis 88:S75–S83.

113. Berthel S. Tuberculosis Drug Accelerator.

114. Dang A. Lilly TB Drug Discovery Initiative. IDRI.

115. More Medicines For Tuberculosis (MM4TB). http://mm4tb.org/.

55 116. Wehrli W, Staehelin M. 1971. Actions of the rifamycins. Bacteriol Rev 35:290–309.

117. Tiberi S, Plessis N du, Walzl G, Vjecha MJ, Rao M, Ntoumi F, Mfinanga S, Kapata N,

Mwaba P, McHugh TD, Ippolito G, Migliori GB, Maeurer MJ, Zumla A. 2018.

Tuberculosis: progress and advances in development of new drugs, treatment regimens, and

host-directed therapies. Lancet Infect Dis 0.

118. Munsiff SS, Kambili C, Ahuja SD. 2006. Rifapentine for the Treatment of Pulmonary

Tuberculosis. Clin Infect Dis 43:1468–1475.

119. Gay JD, DeYoung DR, Roberts GD. 1984. In vitro activities of norfloxacin and

ciprofloxacin against Mycobacterium tuberculosis, M. avium complex, M. chelonei, M.

fortuitum, and M. kansasii. Antimicrob Agents Chemother 26:94–96.

120. Yano T, Kassovska-Bratinova S, Teh JS, Winkler J, Sullivan K, Isaacs A, Schechter

NM, Rubin H. 2011. Reduction of Clofazimine by Mycobacterial Type 2 NADH:Quinone

Oxidoreductase. J Biol Chem 286:10276–10287.

121. Zhang D, Lu Y, Liu K, Liu B, Wang J, Zhang G, Zhang H, Liu Y, Wang B, Zheng M,

Fu L, Hou Y, Gong N, Lv Y, Li C, Cooper CB, Upton AM, Yin D, Ma Z, Huang H. 2012.

Identification of less lipophilic riminophenazine derivatives for the treatment of drug-

resistant tuberculosis. J Med Chem 55:8409–8417.

122. Ford CW, Hamel JC, Wilson DM, Moerman JK, Stapert D, Yancey RJ, Hutchinson DK,

Barbachyn MR, Brickner SJ. 1996. In vivo activities of U-100592 and U-100766, novel

oxazolidinone antimicrobial agents, against experimental bacterial infections. Antimicrob

Agents Chemother 40:1508–1513.

56 123. /LYHUPRUH '0 /LQH]ROLG LQ YLWURௗ PHFKDQLVP DQG DQWLEDFWHULDO VSHFWUXP J

Antimicrob Chemother 51:ii9-ii16.

124. Lippe B von der, Sandven P, Brubakk O. 2006. Efficacy and safety of linezolid in

multidrug resistant tuberculosis (MDR-TB)—a report of ten cases. J Infect 52:92–96.

125. )ORUHV $5 3DUVRQV /0 3DYHOND *HQHWLF DQDO\VLV RI WKH ȕ-lactamases of

Mycobacterium tuberculosis and Mycobacterium smegmaWLVDQGVXVFHSWLELOLW\WRȕ-lactam

antibiotics. Microbiology 151:521–532.

126. Hugonnet J-E, Blanchard JS. 2007. Irreversible Inhibition of the Mycobacterium

WXEHUFXORVLVȕ-lactamase by Clavulanate. Biochemistry (Mosc) 46:11998–12004.

127. Sirturo (bedaquiline) product insert. Silver Spring, MD: Food and Drug

Administration.

128. Andries K, Verhasselt P, Guillemont J, Göhlmann HWH, Neefs J-M, Winkler H, Gestel

JV, Timmerman P, Zhu M, Lee E, Williams P, Chaffoy D de, Huitric E, Hoffner S, Cambau

E, Truffot-Pernot C, Lounis N, Jarlier V. 2005. A Diarylquinoline Drug Active on the ATP

Synthase of Mycobacterium tuberculosis. Science 307:223–227.

129. Koul A, Dendouga N, Vergauwen K, Molenberghs B, Vranckx L, Willebrords R, Ristic

Z, Lill H, Dorange I, Guillemont J, Bald D, Andries K. 2007. Diarylquinolines target

subunit c of mycobacterial ATP synthase. Nat Chem Biol 3:323.

130. Rao SPS, Alonso S, Rand L, Dick T, Pethe K. 2008. The protonmotive force is required

for maintaining ATP homeostasis and viability of hypoxic, nonreplicating Mycobacterium

tuberculosis. Proc Natl Acad Sci 105:11945–11950.

57 131. Koul A, Vranckx L, Dhar N, Göhlmann HWH, Özdemir E, Neefs J-M, Schulz M, Lu

P, Mørtz E, McKinney JD, Andries K, Bald D. 2014. Delayed bactericidal response of

Mycobacterium tuberculosis to bedaquiline involves remodelling of bacterial metabolism.

Nat Commun 5.

132. Huitric E, Verhasselt P, Koul A, Andries K, Hoffner S, Andersson DI. 2010. Rates and

Mechanisms of Resistance Development in Mycobacterium tuberculosis to a Novel

Diarylquinoline ATP Synthase Inhibitor. Antimicrob Agents Chemother 54:1022–1028.

133. Hartkoorn RC, Uplekar S, Cole ST. 2014. Cross-Resistance between Clofazimine and

Bedaquiline through Upregulation of MmpL5 in Mycobacterium tuberculosis. Antimicrob

Agents Chemother 58:2979–2981.

134. Andries K, Villellas C, Coeck N, Thys K, Gevers T, Vranckx L, Lounis N, Jong BC de,

Koul A. 2014. Acquired Resistance of Mycobacterium tuberculosis to Bedaquiline. PLOS

ONE 9:e102135.

135. Matsumoto M, Hashizume H, Tomishige T, Kawasaki M, Tsubouchi H, Sasaki H,

Shimokawa Y, Komatsu M. 2006. OPC-67683, a Nitro-Dihydro-Imidazooxazole

Derivative with Promising Action against Tuberculosis In Vitro and In Mice. PLOS Med

3:e466.

136. Stover CK, Warrener P, VanDevanter DR, Sherman DR, Arain TM, Langhorne MH,

Anderson SW, Towell JA, Yuan Y, McMurray DN, Kreiswirth BN, Barry CE, Baker WR.

2000. A small-molecule nitroimidazopyran drug candidate for the treatment of tuberculosis.

Nature 405:962–966.

137. Manjunatha U, Boshoff HI, Barry CE. 2009. The mechanism of action of PA-824.

Commun Integr Biol 2:215–218.

58 138. Manjunatha UH, Boshoff H, Dowd CS, Zhang L, Albert TJ, Norton JE, Daniels L, Dick

T, Pang SS, Barry CE. 2006. Identification of a nitroimidazo-oxazine-specific protein

involved in PA-824 resistance in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A

103:431–436.

139. Singh R, Manjunatha U, Boshoff HIM, Ha YH, Niyomrattanakit P, Ledwidge R, Dowd

CS, Lee IY, Kim P, Zhang L, Kang S, Keller TH, Jiricek J, Barry CE. 2008. PA-824 Kills

Nonreplicating Mycobacterium tuberculosis by Intracellular NO Release. Science

322:1392–1395.

140. Haver HL, Chua A, Ghode P, Lakshminarayana SB, Singhal A, Mathema B, Wintjens

R, Bifani P. 2015. Mutations in Genes for the F420 Biosynthetic Pathway and a

Nitroreductase Enzyme Are the Primary Resistance Determinants in Spontaneous In Vitro-

Selected PA-824-Resistant Mutants of Mycobacterium tuberculosis. Antimicrob Agents

Chemother 59:5316–5323.

141. Fujiwara M, Kawasaki M, Hariguchi N, Liu Y, Matsumoto M. 2018. Mechanisms of

resistance to delamanid, a drug for Mycobacterium tuberculosis. Tuberculosis 108:186–

194.

142. 2014. Otsuka Wins European Marketing Authorization for DeltybaTM (delamanid).

143. Protopopova M, Hanrahan C, Nikonenko B, Samala R, Chen P, Gearhart J, Einck L,

Nacy CA. 2005. Identification of a new antitubercular drug candidate, SQ109, from a

combinatorial library of 1,2-ethylenediamines. J Antimicrob Chemother 56:968–974.

144. Tahlan K, Wilson R, Kastrinsky DB, Arora K, Nair V, Fischer E, Barnes SW, Walker

JR, Alland D, Barry CE, Boshoff HI. 2012. SQ109 Targets MmpL3, a Membrane

59 Transporter of Trehalose Monomycolate Involved in Mycolic Acid Donation to the Cell

Wall Core of Mycobacterium tuberculosis. Antimicrob Agents Chemother 56:1797–1809.

145. Grzegorzewicz AE, Pham H, Gundi VAKB, Scherman MS, North EJ, Hess T, Jones V,

Gruppo V, Born SEM, Korduláková J, Chavadi SS, Morisseau C, Lenaerts AJ, Lee RE,

McNeil MR, Jackson M. 2012. Inhibition of mycolic acid transport across the

Mycobacterium tuberculosis plasma membrane. Nat Chem Biol 8:334–341.

146. Li K, Schurig-Briccio LA, Feng X, Upadhyay A, Pujari V, Lechartier B, Fontes FL,

Yang H, Rao G, Zhu W, Gulati A, No JH, Cintra G, Bogue S, Liu Y-L, Molohon K, Orlean

P, Mitchell DA, Freitas-Junior L, Ren F, Sun H, Jiang T, Li Y, Guo R-T, Cole ST, Gennis

RB, Crick DC, Oldfield E. 2014. Multitarget Drug Discovery for Tuberculosis and Other

Infectious Diseases. J Med Chem 57:3126–3139.

147. Makarov V, Manina G, Mikusova K, Mollmann U, Ryabova O, Saint-Joanis B, Dhar

N, Pasca MR, Buroni S, Lucarelli AP, Milano A, De Rossi E, Belanova M, Bobovska A,

Dianiskova P, Kordulakova J, Sala C, Fullam E, Schneider P, McKinney JD, Brodin P,

Christophe T, Waddell S, Butcher P, Albrethsen J, Rosenkrands I, Brosch R, Nandi V,

Bharath S, Gaonkar S, Shandil RK, Balasubramanian V, Balganesh T, Tyagi S, Grosset J,

Riccardi G, Cole ST. 2009. Benzothiazinones Kill Mycobacterium tuberculosis by

Blocking Arabinan Synthesis. Science 324:801–804.

148. Makarov V, Lechartier B, Zhang M, Neres J, van der Sar AM, Raadsen SA, Hartkoorn

RC, Ryabova OB, Vocat A, Decosterd LA, Widmer N, Buclin T, Bitter W, Andries K, Pojer

F, Dyson PJ, Cole ST. 2014. Towards a new combination therapy for tuberculosis with next

generation benzothiazinones. EMBO Mol Med 6:372–383.

60 149. Pasca MR, Degiacomi G, Ribeiro AL d. JL, Zara F, De Mori P, Heym B, Mirrione M,

Brerra R, Pagani L, Pucillo L, Troupioti P, Makarov V, Cole ST, Riccardi G. 2010. Clinical

Isolates of Mycobacterium tuberculosis in Four European Hospitals Are Uniformly

Susceptible to Benzothiazinones. Antimicrob Agents Chemother 54:1616–1618.

150. Mikusova K, Huang H, Yagi T, Holsters M, Vereecke D, D’Haeze W, Scherman MS,

Brennan PJ, McNeil MR, Crick DC. 2005. Decaprenylphosphoryl Arabinofuranose, the

Donor of the D-Arabinofuranosyl Residues of Mycobacterial Arabinan, Is Formed via a

Two-Step Epimerization of Decaprenylphosphoryl Ribose. J Bacteriol 187:8020–8025.

151. Trefzer C, Rengifo-Gonzalez M, Hinner MJ, Schneider P, Makarov V, Cole ST,

Johnsson K. 2010. Benzothiazinones: Prodrugs That Covalently Modify the

Decaprenylphosphoryl-ȕ-D-ribose 2’-epimerase DprE1 of Mycobacterium tuberculosis. J

Am Chem Soc 132:13663–13665.

152. Trefzer C, Škovierová H, Buroni S, Bobovská A, Nenci S, Molteni E, Pojer F, Pasca

MR, Makarov V, Cole ST, Riccardi G, Mikušová K, Johnsson K. 2012. Benzothiazinones

Are Suicide Inhibitors of Mycobacterial Decaprenylphosphoryl-ȕ- d-ribofuranose 2’-

Oxidase DprE1. J Am Chem Soc 134:912–915.

153. Neres J, Pojer F, Molteni E, Chiarelli LR, Dhar N, Boy-Rottger S, Buroni S, Fullam E,

Degiacomi G, Lucarelli AP, Read RJ, Zanoni G, Edmondson DE, De Rossi E, Pasca MR,

McKinney JD, Dyson PJ, Riccardi G, Mattevi A, Cole ST, Binda C. 2012. Structural Basis

for Benzothiazinone-Mediated Killing of Mycobacterium tuberculosis. Sci Transl Med

4:150ra121-150ra121.

61 154. Batt SM, Jabeen T, Bhowruth V, Quill L, Lund PA, Eggeling L, Alderwick LJ, Futterer

K, Besra GS. 2012. Structural basis of inhibition of Mycobacterium tuberculosis DprE1 by

benzothiazinone inhibitors. Proc Natl Acad Sci 109:11354–11359.

155. Lechartier B, Hartkoorn RC, Cole ST. 2012. In Vitro Combination Studies of

Benzothiazinone Lead Compound BTZ043 against Mycobacterium tuberculosis.

Antimicrob Agents Chemother 56:5790–5793.

156. Pethe K, Bifani P, Jang J, Kang S, Park S, Ahn S, Jiricek J, Jung J, Jeon HK, Cechetto

J, Christophe T, Lee H, Kempf M, Jackson M, Lenaerts AJ, Pham H, Jones V, Seo MJ, Kim

YM, Seo M, Seo JJ, Park D, Ko Y, Choi I, Kim R, Kim SY, Lim S, Yim S-A, Nam J, Kang

H, Kwon H, Oh C-T, Cho Y, Jang Y, Kim J, Chua A, Tan BH, Nanjundappa MB, Rao SPS,

Barnes WS, Wintjens R, Walker JR, Alonso S, Lee S, Kim J, Oh S, Oh T, Nehrbass U, Han

S-J, No Z, Lee J, Brodin P, Cho S-N, Nam K, Kim J. 2013. Discovery of Q203, a potent

clinical candidate for the treatment of tuberculosis. Nat Med 19:1157–1160.

157. Cook GM, Hards K, Vilchèze C, Hartman T, Berney M. 2014. Energetics of Respiration

and Oxidative Phosphorylation in Mycobacteria. Microbiol Spectr 2.

158. Kalia NP, Hasenoehrl EJ, Rahman NBA, Koh VH, Ang MLT, Sajorda DR, Hards K,

Grüber G, Alonso S, Cook GM, Berney M, Pethe K. 2017. Exploiting the synthetic lethality

between terminal respiratory oxidases to kill Mycobacterium tuberculosis and clear host

infection. Proc Natl Acad Sci 201706139.

62 Chapter 2: Arylvinylpiperazine amides, a new class of potent inhibitors targeting QcrB of Mycobacterium tuberculosis

Caroline S. Foo1§, Andréanne Lupien1§, Maryline Kienle2, Anthony Vocat1, Andrej Benjak1, Raphael Sommer1, Dirk A. Lamprecht3#, Adrie JC Steyn3,4, Kevin Pethe5, Jérémie Piton1, Karl-Heinz Altmann2,Stewart T. Cole1#

1Global Health Institute, École Polytechnique Fédérale de Lausanne, Switzerland

2Institute of Pharmaceutical Sciences, Department of Chemistry and Applied Biosciences, ETH

Zürich, Switzerland

3Africa Health Research Institute, South Africa

4Department of Microbiology, University of Alabama at Birmingham, Birmingham, AL, United

States

5 Lee Kong Chian School of Medicine and School of Biological Sciences, Nanyang Technological

University, Singapore

§These authors contributed equally to the design and execution of experiments

# current affiliations: D.A.L. - Janssen Pharmaceutica, Turnhoutweg 30, 2340, Beerse, Belgium;

S.T.C. – Institut Pasteur, Paris, France

Manuscript accepted, mBio, August 2018

Contributions: design and execution of experiments, data analysis, manuscript preparation

ABSTRACT

New drugs are needed to control the current tuberculosis (TB) pandemic caused by infection with

Mycobacterium tuberculosis. We report here on our work with AX-35, an arylvinylpiperazine amide and four related analogs, which are potent anti-tubercular agents in vitro. All five compounds showed good activity against M. tuberculosis in vitro and in infected THP-1 macrophages, while displaying only mild cytotoxicity. Isolation and characterization of M. tuberculosis resistant mutants to the arylvinylpiperazine amide derivative AX-35 revealed mutations in the qcrB gene encoding a subunit of cytochrome bc1 oxidase, one of two terminal oxidases of the electron transport chain. Cross- resistance studies, allelic exchange, transcriptomic analyses and bioenergetics flux assays provided conclusive evidence that the cytochrome bc1-aa3 is the target of AX-35, although the compound appears to interact differently with the quinol binding pocket compared to previous QcrB inhibitors.

The transcriptomic and bioenergetic profiles of M. tuberculosis treated with AX-35 were similar to those generated by other cytochrome bc1 oxidase inhibitors, including the compensatory role of the alternate terminal oxidase cytochrome bd in respiratory adaptation. In the absence of cytochrome bd oxidase, AX-35 was bactericidal against M. tuberculosis. Finally, AX-35 and its analogs were active in an acute mouse model of TB infection, with two analogs displaying improved activity over the parent compound. Our findings will guide future lead optimization to produce a drug candidate for the treatment of TB and other mycobacterial diseases, including Buruli ulcer and leprosy. (235 words)

65 IMPORTANCE

New drugs against Mycobacterium tuberculosis are urgently needed to deal with the current global TB pandemic. We report here on the discovery of a series of arylvinylpiperazine amides (AX-35 – AX-

39), which represent a promising new family of compounds with potent in vitro and in vivo activity against M. tuberculosis. AX compounds target the QcrB subunit of the cytochrome bc1 terminal oxidase with a different mode of interaction compared to known QcrB inhibitors. This study provides the first multi-faceted validation of QcrB inhibition by recombineering-mediated allelic exchange, gene expression profiling and bioenergetic flux studies. It also provides further evidence for the compensatory role of cytochrome bd oxidase upon QcrB inhibition. In the absence of cytochrome bd oxidase, AX compounds are bactericidal, an encouraging property for future antimycobacterial drug development.

KEYWORDS tuberculosis, QcrB inhibitor, cytochrome bc1 oxidase, cytochrome bd oxidase, mycobacterial respiration, mycobacterial diseases

66 INTRODUCTION

Mycobacterium tuberculosis, the causative agent of tuberculosis (TB), is at the origin of a severe global health problem. TB is the leading cause of death due to an infectious agent, with an estimated death toll of 1.6 million in 2016 (1). Drug-susceptible TB is typically treated over the course of six months with a combination of the first-line drugs isoniazid, rifampicin, pyrazinamide and ethambutol; however, this regimen has been rendered ineffective nowadays in many cases, due to the emergence of drug-resistant TB (1). Treatment of multi-drug resistant TB (MDR-TB), whose defining characteristics are resistance to both rifampicin and isoniazid, requires a regimen of second-line drugs including a fluoroquinolone and injectable agents, but treatment success rates are lower than 50% (1–

3). In addition to resistance to rifampicin and isoniazid, extensively-drug resistant TB (XDR-TB) strains are resistant to at least one of the fluoroquinolones and an injectable. XDR-TB cure rates are low at 30% (1). Long treatment duration, toxicity issues, drug resistance and the expense of the second-line regimens are important factors driving the search for new anti-TB drugs with novel mechanisms of action.

To generate new anti-TB candidates, drug screening efforts have employed phenotypic or whole-cell screening approaches recently, whereby compound libraries of new chemical entities are screened against M. tuberculosis for inhibition of bacterial growth in various settings which, ideally, should recapitulate pathophysiological conditions (4–6). Attempts are subsequently made to identify and validate the target(s) of hits to enable and facilitate further compound optimization. The diarylquinoline bedaquiline (BDQ), which targets the subunit c of mycobacterial ATP synthase (7, 8), emerged from such a compound-to-target approach and has been hailed as a milestone for TB drug discovery and, with its approval in 2012, became the first anti-TB drug to be approved in more than

40 years. The imidazopyrimidine amide Q203 and lansoprazole sulfide (LPZS) both target the QcrB subunit of cytochrome bc1 oxidase and were identified from intracellular screens with M. tuberculosis- infected macrophages and fibroblasts, respectively (9, 10). The approval of BDQ for the treatment of

67 MDR-TB, inclusion of BDQ as a component of the highly promising NIX-TB combination trial for

XDR-TB treatment (11), and the fact that Q203 is currently in Phase I clinical development (2, 12), altogether validate mycobacterial respiration as a new, relevant and attractive target for an obligate aerobic pathogen.

In 2013, GlaxoSmithKline published the results of a large phenotypic screening campaign against

Mycobacterium bovis BCG and M. tuberculosis, which yielded a total of 177 low-molecular weight hits (13). One of these compounds was GW861072X (referred to here as AX-35) (Fig. 1), which exhibited potent activity against both M. bovis BCG and M. tuberculosis H37Rv (MIC of 0.3 μM against both species).

Due to its structural simplicity, we considered AX-35 as an attractive starting point for lead optimization and prepared a series of analogs for structure-activity-relationship (SAR) studies, the full details of which will be published in a forthcoming report. In this paper, we describe the characterization of AX-35 and four of the most potent analogs that have emerged from our SAR work

(AX-36 – AX-39) and are active against M. tuberculosis replicating in vitro and in infected macrophages. Subsequent work validated the target of these compounds as QcrB, the b subunit of the cytochrome bc1 oxidase, based on the isolation of resistant mutants, cross-resistance studies, RNA-seq and bioenergetics flux assays. Finally, further characterization of three AX compounds in an acute mouse model of TB demonstrated in vivo efficacy against M. tuberculosis.

RESULTS

Initial characterization of AX-35 and its analogs. The synthesis of AX-35 – AX-39 is described in the supporting information (for structures see Table 1). Analog AX-36 served to assess the importance of the position of the sulfur atom in the thiophene ring for antimycobacterial activity, while thiazole derivatives AX-37 and AX-39 were designed to probe the effect of enhanced compound polarity. In addition, replacement of the thiophene ring by a thiazole moiety could mitigate the toxicity

68 risk often associated with thiophene derivatives. For the same reason, the thiophene moiety in AX-35 was also replaced by a phenyl ring, as an established bioisostere (AX-38). As shown in the data in

Table 1, AX-35 is the most active compound within this series, with an MIC of 0.05 μg/ml against M. tuberculosis H37Rv; AX-36, AX-37, AX-38, and AX-39 have somewhat higher MIC values ranging from 0.1 to 0.3 μg/ml (Table 1). The activity of AX compounds is selective for pathogenic mycobacteria in the panel of bacterial strains and fungi tested (Table 2). Members of the M. tuberculosis complex were especially susceptible, notably Mycobacterium canettii, followed by

Mycobacterium ulcerans. Most of the microorganisms tested were resistant to both AX-35 and AX-

36. Mild to no cytotoxicity was observed in human HepG2 cells, resulting in selectivity indices

(HepG2 TD50/MIC) above 250 for all compounds, with AX-35 being the most selective (Table 1). The

IC50 values for M. tuberculosis H37Rv-infected THP-1 macrophages ranged from 0.1 to 1.8 μg/ml

(Table 1), reflecting the potent ex vivo activity of these molecules. The metabolic stability of the compounds in mouse and human liver microsomes was moderate to low (Table S1).

Evidence of AX compounds targeting QcrB of M. tuberculosis. Initial attempts to raise spontaneous resistant mutants in M. tuberculosis H37Rv to AX-35 and AX-36 on solid 7H10 medium failed, even at exposures of up to 100x MIC. An alternative method of continual passaging of H37Rv in complete 7H9 liquid medium containing AX-35 or AX-36 at concentrations above the MIC, beginning at 2x MIC and increasing to 100x MIC over five passages, proved to be more successful.

Increasing MICs reflected the gradual selection of a resistant sub-population in the culture on constant exposure to the compounds, especially after the fourth and fifth passages. Single clones isolated on

7H10 plates at passage five were subsequently tested for resistance to AX-35 or AX-36 by MIC determinations using the resazurin microtiter assay plate (REMA) method (Fig. 2a).

To identify mutations associated with resistance to AX-35 or AX-36, whole-genome sequencing

(WGS) analysis was performed on six clones with a 50-fold increase in MIC compared to the parental strain H37Rv (Fig. 2a). Three missense mutations in qcrB were revealed, leading to non-synonymous

69 substitutions in QcrB, the b subunit of cytochrome bc1 oxidase. S182P and M342V mutations were associated with AX-35 resistance and M342I with AX-36 resistance. The specific contribution of these substitutions to AX-35 and AX-36 resistance was confirmed by the generation of the same missense mutations in chromosomal qcrB of wild-type H37Rv by recombineering and MIC determination (Fig.

S1).

Since mutations in qcrB have also been identified in mutants of M. tuberculosis H37Rv resistant to

Q203 and LPZS, two compounds known to target QcrB, cross-resistance studies were performed to investigate if AX-35 and AX-36 behaved in a similar manner. Q203- and LPZS-resistant strains, harbouring single-nucleotide polymorphisms leading to T313A and L176P in QcrB, respectively, are cross-resistant to AX-35 and AX-36 (Fig. 2b). These results indicate that the binding mode of the AX compound with QcrB shares some similarities with Q203 and LPZS. However, not all AX-resistant mutants were fully resistant to Q203 and LPZS (Fig. 2b), thus suggesting a distinct mode of binding for these QcrB inhibitors. On inspection of a model of M. tuberculosis QcrB, residues S182, M342,

T313 and L176 were found to cluster around the quinol oxidation site of the enzyme (Fig. 2c), implying that AX-35 and AX-36 bind to QcrB at this pocket, similar to Q203 and LPZS.

As QcrB is one of the respiratory subunits of the cytochrome bc1 oxidase, a consequence of its inhibition is depletion of ATP levels in the bacterium (9, 10). The intracellular ATP level was measured in wild-type H37Rv after 24 h exposure to AX-35 or AX-36 and found to be ~90% lower than for the untreated control, similar to the effect of treatment with the ATP-synthase inhibitor BDQ or with Q203 under the same conditions (Fig. 2d), thus indicating that AX compounds do indeed affect ATP levels.

ATP was also depleted in BDQ- or Q203-exposed AX-resistant mutants, with ATP levels similar to that in BDQ- or Q203-exposed wild-type M. tuberculosis. However, ATP levels were essentially unaffected in the AX-resistant strains in the presence of AX-35 or AX-36 (Fig. 2d). This demonstrates that the mutations in QcrB associated with AX-resistance do not greatly impact the activity of BDQ nor Q203, indicating that AX compounds do not have off-target effects on the ATP synthase and that

70 they interact in a different manner with QcrB compared to Q203. Taken together these findings indicate that QcrB is the direct target of the AX compounds.

Transcriptional response of M. tuberculosis to AX-35 treatment. To gain insight into the initial adaptive response of M. tuberculosis to AX-35 treatment, the transcriptomes of wild-type

H37Rv exposed to AX-35 at 10x and 30x MIC for a duration of 4h were examined. Of 81 genes significantly down-regulated (fold-change (FC) -2, Benjamini and Hochberg’s adjusted p-value (padj)

leuC and leuD (involved in leucine biosynthesis) were the most extensively down-regulated

(Fig. 3a, b). Thirteeen genes were significantly up-regulated (FC 2, padj LQWKHSUHVHQFHRI

AX-35, seven of which were involved in intermediary metabolism and respiration (Fig. 3a, b). The most highly up-regulated gene was lipU, a gene implicated in lipid hydrolysis. Notably, two main operons in the M. tuberculosis transcriptome were up-regulated, namely the cyd and mymA operons.

The cyd operon consists of cydA, cydB and cydD (cydA and cydB were close to the cutoff with FC 1.97 and 1.99 respectively), with cydA and cydB coding for subunits I and II, respectively, of the cytochrome bd oxidase, while cydDC encodes an ABC transporter involved in cytochrome bd assembly. The mymA operon from Rv3083 to Rv3089 (14) includes tgs4 and rv3088 encoding triacylglycerol synthases (15) lipR, rv3085 and rv3086 encoding short-chain dehydrogenases, and an acyl-CoA synthase gene, fadD13.

To validate the RNA-seq data, qRT-PCR was performed targeting lipU, cydB, and tgs4 in both wild- type H37Rv and its AX-resistant mutants after exposure to AX-35 at 30x MIC (Fig. 3c). The up- regulation of all three genes in wild-type H37Rv upon AX-35 treatment detected by qRT-PCR is consistent with the RNA-seq data, with lipU having the highest relative expression of around 6-fold compared to the control, i.e. H37Rv treated with DMSO alone. No significant differences in the expression levels of these three genes were measured for the AX-resistant mutant when treated with

DMSO or with AX-35, therefore establishing that their up-regulation and, by extension, the de-

71 regulation of genes observed by RNA-seq, is indeed a consequence of AX-35 treatment in M. tuberculosis.

Respiratory response of M. tuberculosis to AX-35 treatment. Bioenergetic flux studies were carried out to assess the respiratory response of M. tuberculosis to AX-35, with Q203 used as a control.

Oxygen consumption rate (OCR) was first measured in M. tuberculosis H37Rv in the presence of glucose to determine the basal OCR and the subsequent addition of AX-35 or Q203 resulted in an increase in OCR (Fig. 4a, b). Lastly, the protonophore carbonyl cyanide m-chlorophenyl hydrazine

(CCCP) was added, depolarising the bacterial membrane and enabling the measurement of the maximum respiration capability of the bacteria.

The increase in OCR upon the addition of AX-35 or Q203 can be attributed to cytochrome bd oxidase, since this increase was no longer observed in the presence of either compound in the ǻcydAB strain

(Fig. 4a, b). This switch in M. tuberculosis respiration to the bd oxidase has previously been validated as a respiratory signature of cytochrome bc1 oxidase inhibitors such as Q203 (16), and our results indicate a similar bioenergetic profile of AX-35 to that of Q203.

Interestingly, a difference can be observed in the bioenergetic profiles, on exposure of the cytochrome bd oxidase KO strain harbouring a QcrB A317V mutation to these two QcrB inhibitors. Addition of

AX-35 decreases the OCR of this strain, generating a profile similar to that of H37RvǻcydAB (Fig.

4a), implying that respiration is via the bd oxidase branch and that QcrB remains fully inhibited despite the A317V mutation. The basal OCR remains unaffected by the addition of Q203, however (Fig. 4b), which would suggest a partial inhibition of QcrB and thus inability of the drug to fully inhibit QcrB due to the mutation. This therefore reveals additional differences in the binding modes of AX-35 and

Q203 to QcrB.

Taken together, these results demonstrate the capability of respiratory adaptation of M. tuberculosis via the bd oxidase upon treatment with AX-35, generating a respiratory signature similar to that of

72 another established QcrB inhibitor, Q203, albeit highlighting different modes of interaction of the two compounds within the binding pocket.

Cidality, drug interactions, and in vivo efficacy of AX compounds. QcrB inhibitors described to date are bacteriostatic due to the compensatory role of the cytochrome bd oxidase. To determine the mode of action of AX compounds against M. tuberculosis, minimum bactericidal concentrations (MBC) were determined in M. tuberculosis +5Y +5YǻcydAB, and complemented H37RvǻcydAB::cydAB strains (Fig. S2). In H37Rv, AX-35 is bacteriostatic with an

MBC up to a concentration 32-fold higher than the MIC. Conversely, AX-35 was bactericidal against the +5YǻcydAB strain at a concentration 4-fold in excess of its MIC. Cidality of AX-35 can be specifically attributed to the lack of the cytochrome bd oxidase, as evident from the survival of the bacteria in the complemented strain treated with AX-35.

Since synergistic interactions have been previously reported between the cell wall inhibitor PBTZ169 and BDQ (17), as well as between compounds targeting the mycobacterial respiratory chain (16, 18), checkerboard assays for two-drug combinations were performed to determine the nature of interactions between AX-35 with PBTZ169, BDQ, or clofazimine (CFZ) in M. tuberculosis H37Rv. The Ȉ),& indices obtained for all three combinations range from 0.8 to 1.6 for concentrations of AX-35 less than or equal to 0.5-fold MIC (Table S2), indicating non-antagonistic, additive interactions.

The activity of AX compounds in vivo was determined in mouse models of chronic and acute TB.

None of the AX compounds reduced the bacterial burden in mice at the dose of 100 mg/kg in the chronic model (Fig. S3). In the acute model, however, AX-35 at an oral dose of 200 mg/kg, AX-37 and AX-39 at 100 mg/kg significantly reduced the bacterial load in mouse lungs by 0.4 log10, 0.6 log10 and 0.9log10, respectively, compared to the TPGS vehicle control (Fig. 5).

73 DISCUSSION

The compounds investigated in this study are chemically distinct and different from the diverse QcrB inhibitors described in recent years (9, 19–25). Of the three mutations in QcrB identified here in spontaneous mutants resistant to AX-35 and AX-36, namely S182P, M342V, and M342I, the first two have been previously reported (20). Cross-resistance studies and bioenergetic flux assays highlight influential residues for the interaction of AX-35 and AX-36 with QcrB, which occurs in the same binding site as Q203 and LPZS. In particular, mutations S182P, M342V/I, T313A, L176P, impair the interaction of AX-35 and AX-36 with QcrB, resulting in resistance of M. tuberculosis to the compounds. The substitution A317V, however, does not interfere with this interaction. This highlights the different modes of interaction of AX-35 and AX-36 with QcrB compared with Q203 or LPZS, and this information will be useful for further optimization of the compounds through structure-based drug design.

In addition to cross-resistance studies, transcriptomic and bioenergetic flux studies provided further evidence of QcrB as the primary target of AX-35/36. Treatment of M. tuberculosis with AX-35 results in the up-regulation of the cydABD operon, but not the dosR regulon, which is consistent with the transcriptomic signature of respiratory inhibitors specifically inhibiting the cytochrome bc1-aa3 terminal oxidase (26). In line with this, exposure of M. tuberculosis to AX-35 results in an increase in

OCR due to activity of the cytochrome bd oxidase, generating a bioenergetic profile similar to that of the respiratory signature of cytochrome bc1 oxidase inhibitors such as Q203 (16).

Cytochrome bd oxidase is non-proton translocating and thus the less bioenergetically efficient of the two terminal oxidases of M. tuberculosis (27). It has been demonstrated that the bd oxidase is up- regulated under conditions where the function of the cytochrome bc1-aa3 is compromised (20, 28–30).

In addition, inhibitors of respiration such as BDQ and QcrB inhibitors have markedly improved activity upon deletion or inhibition of bd oxidase (20, 31–33). The pronounced compensatory role of this alternate oxidase in the respiratory adaptation of M. tuberculosis for its survival is also apparent

74 in our study, where the absence of bd oxidase is associated with AX-mediated bactericidal activity.

Targeting both terminal oxidases simultaneously may therefore be a novel and effective strategy against M. tuberculosis. Although the increase in respiration, up-regulation of cytochrome bd oxidase,

DQGHQKDQFHGNLOOLQJRIǻcydAB mutants upon inhibition of ATP synthase by BDQ in M. smegmatis

(38) have been attributed to the ionophoric/off-target effect of BDQ (39), it is unlikely that AX compounds behave in a similar manner as uncouplers, based on the bioenergetic flux data.

It is also of note that AX treatment of M. tuberculosis results in the up-regulation of tgs genes (tgs4 and rv3087) and lipases (lipU and lipR). tgs genes are involved in the biosynthesis of triacylglycerol

(TAG), the predominant energy source for M. tuberculosis in host macrophages (15, 34), while lip genes encode lipolytic enzymes thought to be involved in the utilization of host TAG (35). Such up- regulation of tgs and lip genes has been observed in various models of host stress (36, 37), and similarly

AX-induced stress re-models central carbon metabolism in M. tuberculosis toward lipid metabolism.

The compounds investigated here show promising properties as anti-TB agents. AX-35 to AX-39 demonstrated potent activity against M. tuberculosis in vitro and in macrophages, while having mild to no cytotoxicity. When tested in an acute mouse model of TB, AX-35, AX-37 and AX-39 were capable of reducing the bacterial burden in the lungs of mice. This in vivo activity can likely be improved further by addressing metabolic stability issues, such as those observed in mouse microsomes, although the compounds appear to be more stable in human microsomes.

The fact that activity of the AX compounds was only observed in the acute rather than the chronic mouse model of TB could be attributed to the difference in expression levels of the terminal oxidases found between these two stages of M. tuberculosis lung infection (29). The cytochrome bc1-aa3 oxidase is more highly expressed than the bd oxidase during the exponential phase of growth of M. tuberculosis in mice, whereas it is down-regulated as a response to host immunity and remains down- regulated as the bacteria enter a non-replicating state in the chronic phase (29). Thus, targeting QcrB

75 during the exponential phase of M. tuberculosis growth would have a bigger impact on bacterial viability despite any compensatory response by bd oxidase. Another potential application of the AX- inhibitors could be in the treatment of the human diseases Buruli ulcer and leprosy as their causative agents, M. ulcerans and M. leprae, respectively, lack the cytochrome bd oxidase.

In conclusion, our SAR data prove that structural changes to AX-35 (GW861072X) can lead to improved in vivo activity against M. tuberculosis and that the arylvinylpiperazine amide series target the QcrB subunit of the cytochrome bc1-aa3 oxidase. As part of the mycobacterial respiratory chain,

QcrB is an attractive target for combination therapy. Furthermore, the biochemical information obtained from this study could guide future lead optimization work to enhance potency and stability of the compounds.

76 MATERIALS AND METHODS

Drugs used in this study. Q203 was synthesized as described before (9, 40), LPZS was purchased from Santa Cruz biotechnology, Toronto Research Chemicals Inc., BDQ was a gift from

Janssen Pharmaceutica NV, PBTZ169 (Macozinone) was provided by Innovative Medicines for

Tuberculosis, while CFZ, INH and RIF were from Sigma Aldrich.

Culture conditions of M. tuberculosis strains, other bacteria and eukaryotic cell lines.

Mycobacterial strains were grown at 37oC in Middlebrook 7H9 broth (Difco) supplemented with 0.2% glycerol, 0.05% Tween-80 and 10% albumin dextrose catalase (ADC) (7H9 complete) or on 7H10 agar plates supplemented with 0.5% glycerol and 10% oleic ADC. Bacillus subtilis, Candida albicans,

Corynebacterium diphtheriae, Corynebacterium glutamicum, Escherichia coli, Micrococcus luteus,

Pseudomonas putida, Salmonella typhimurium,and Staphylococcus aureus were grown in LB broth.

Enterococcus faecalis, Listeria monocytogenes, and Pseudomonas aeruginosa were grown in brain heart infusion (BHI) broth. HepG2 cells were grown in DMEM (Gibco) media supplemented with

o 10% fetal bovine serum at 37 C with 5% CO2. THP-1 macrophages were grown in RPMI medium

o supplemented with 10% fetal bovine serum and 1 mM sodium pyruvate at 37 C with 5% CO2.

Determination of MICs. MICs were determined using the resazurin reduction microplate assay (REMA) as previously described (41). Strains were grown to log-phase (OD600 0.4 to 0.8) and diluted to an OD600 of 0.0001. Two-fold serial dilutions of each test compound were prepared in 96- well plates containing 100 μl of bacteria per well (3 x 103 cells per well). Plates were incubated either at 30oC or 37oC as required with appropriate incubation times (e.g. for M. tuberculosis at 37oC, 6 days).

10 μl of resazurin (0.0025% w/v) was added to each well, after which the fluorescence intensity of the resorufin metabolite (excitation/emission: 560/590 nm) was read using an Infinite F200 Tecan plate reader. MIC values representing 90% growth inhibition were determined by a non-linear fitting of the data to the Gompertz equation using GraphPad Prism.

77 Cytotoxicity for HepG2 cells. Human HepG2 cells (4,000 cells/well) were incubated for 3

o days with two-fold serially diluted compounds at 37 C, under an atmosphere of 5% CO2. Cell viability was determined by the addition of resazurin (0.0025% w/v) for 4 h at 37oC and the fluorescence intensity measured as in REMA.

Assessment of drug activity in THP-1 macrophages. THP-1 human monocytic cells

(105/well) were seeded in 96-well plates and incubated with 100 nM phorbol-12-myristate-13-acetate

(PMA) overnight to stimulate macrophage differentiation. Differentiated macrophages were infected with M. tuberculosis H37Rv grown to log phase (OD600 0.4 to 0.8) at MOI 5. Extracellular bacteria

o were removed after 3-4 h incubation at 37 C with 5% CO2 by removing the RPMI medium and washing with PBS. Compounds to be tested were prepared in separate 96-well plates by two-fold serial dilutions in a final volume of 100 μl RPMI which was then transferred to the plates of infected THP-1

o ® macrophages. Plates were sealed and incubated for 48 h at 37 C with 5% CO2. 10 μlof PrestoBlue

o (ThermoFischer Scientific) was added and plates were incubated for up to 1 h at 37 C with 5% CO2 before the fluorescence intensity (excitation/emission: 560/590 nm) was measured using Infinite F200

Tecan plate reader. Dose-response curves were plotted and IC50 values obtained using a non-linear regression fit equation (log[inhibitor] vs response, variable slope) in GraphPad Prism.

Microsomal stability studies. Metabolic stability of the compounds was measured based on intrinsic clearance (Clint) in mouse and human liver microsomes as previously described (42). Final compound concentration in the mixture of microsomes and NADPH-regeneration system was 2 μg/ml, and a mixture without NADPH-regeneration was also prepared for each compound as a control of the stability of the compound with time. Carbamazepine and nifedipine at 2 μg/ml were used as low and high intrinsic clearance controls, respectively.

Mouse studies. To assess compound activity in an acute mouse model of TB, female BALB/c mice from Charles River Laboratories (7-8 weeks old, 20g, 5 mice per group) were aerosol-infected

78 with M. tuberculosis H37Rv at a low dose and treated by oral gavage the following day with either vehicle controls (TPGS 20% or TPGS 20% containing 1% DMSO) or test compounds once daily for

10 days. AX compounds were prepared in TPGS 20%, Q203 in TPGS 20% containing 1% DMSO (as described in (32) and INH in ddH2O. Compounds were ground with a pestle and mortar followed by sonication at room temperature for 30 min. Compound solutions were stored at 4oC and prepared freshly after 5 days. The day after the final treatment, mice were sacrificed and serial dilutions of lung homogenates were plated on 7H10 agar containing 10 μg/ml cyclohexamide and 50 μg/ml ampicillin.

Experiments were approved by the Swiss Cantonal Veterinary Authority (authorization number 3082).

Isolation and characterization of AX-resistant mutants. AX-resistant mutants of M. tuberculosis H37Rv were isolated from 7H9 cultures over 5 passages with increasing concentrations of AX-35 or AX-36 starting from 2x, 5x, 10x MIC to a final concentration of 50x and 100x MIC.

Single colonies were obtained from three independent cultures by streaking on 7H10 agar plates and resistance to AX was measured by REMA. Genomic DNA extraction was performed using the

QiaAMP UCP Pathogen Minikit (Qiagen) as per the manufacturer’s instructions. Whole genome sequencing was performed using Illumina technology with sequencing libraries prepared using the

KAPA HyperPrep kit (Roche) and sequenced on an Illumina HiSeq 2500 instrument. All raw reads were adapter- and quality-trimmed with Trimmomatic v0.33 (43) and mapped onto the M. tuberculosis

H37Rv reference genome (RefSeq NC_000962.3) using Bowtie2 v2.2.5 (44). The bamleftalign program from the FreeBayes package v0.9.20-18 (45) was used to left-align indels. Reads with mapping quality below 8 and duplicate reads were omitted.

Variant analysis. Variant calling was done using VarScan v2.3.9 (46) using the following cut- offs: minimum overall coverage of ten non-duplicated reads, minimum of five non-duplicated reads supporting the SNP, base quality score >15, and a SNP frequency above 30%. The rather low thresholds, especially the SNP frequency, were deliberately chosen to avoid missing potential variants

79 in alignment-difficult regions, or in case of mixed population. All putative variants unique to the mutant strains were manually checked by inspecting the alignments.

Recombineering for target confirmation. Insertion of mutations associated with AX resistance in chromosomal qcrB of M. tuberculosis was done based on a recombineering method (47).

Desired mutations were centered in lagging strand oligos of 70 nucleotides. M. tuberculosis H37Rv containing plasmid pJV53 was grown to OD600 of 0.8 in 7H9 liquid media containing 25 μg/ml kanamycin and exposed to 0.2% acetamide for 24 h and 0.2 M glycine for 16 h. Competent cells were prepared and co-transformed with 100 ng of lagging oligo and 50 ng of pYub412 carrying a hygromycin selection marker. After a rescue of 3 days at 37°C, cells were plated on 7H10 plates containing hygromycin at 50 μg/ml and transformants were evaluated for resistance to AX by REMA.

Desired mutations associated with AX resistance were subsequently confirmed by Sanger sequencing across qcrB and the absence of other mutations was confirmed by whole-genome sequencing.

Modelling of M. tuberculosis QcrB. The QcrB protein of M. tuberculosis was modelled using the homology modelling web server (SWISS) (48) and the chain A of the crystal structure of the mutant

Rhodobacter sphaeroides cytochrome bc1 oxidase (PDB code: 2QJK). Illustrations were made using

Pymol software (49).

Quantification of intracellular ATP. Log-phase cultures of wild-type H37Rv or AX-resistant mutants (about 106 CFU/ml) were exposed to test compounds at 2.5x MIC for 24 h in a final volume of 100 μl and incubated with BacTiter Glo Reagent (Promega) (v/v 4:1) for 5 min in the dark.

Luminescence was measured on a TECAN Infinite M200 in relative light units (RLU) with an integration time of 1s.

Total RNA extraction and RNA-seq. Wild-type and AX-resistant H37Rv cultures were grown to mid-log phase and exposed to DMSO (vehicle control) or compounds for 4 h at 37oC. Cells were harvested by centrifugation and pellets were stored with 1 ml of TRIzol reagent (Thermo Fisher

80 Scientific) at -80oC until further processed. Cells were lysed by bead-beating and total RNA was extracted by phenol-chloroform with DNAse treatment (RQ1 RNase–free DNase, Promega). Library preparation was done using the Ribo-zero rRNA removal kit (Illumina) for Gram-positive bacteria to deplete rRNA from total RNA. Two biological replicates for each strain were prepared for RNA-seq.

Reads were adapter- and quality-trimmed with Trimmomatic v0.33 (43) and mapped onto the M. tuberculosis H37Rv reference genome (RefSeq NC_000962.3) using Bowtie2 v2.2.5 (44) Counting reads over features was done with featureCounts from the Subread package v1.4.6 (50). DESeq2 (51) as used to infer differentially expressed genes.

Quantitative PCR. cDNA was prepared from total RNA using SuperScript III First-strand

Synthesis kit (Invitrogen) and analysed by qPCR for targeted gene expression in duplicates using

Power SYBR Green PR Master Mix (Applied Biosystems) on a QuantStudio 5 Real-Time PCR system

(Thermo Fisher Scientific). sigA was used as a house-keeping gene for normalization and the ǻǻCt method was used for quantification.

Extracellular Flux Analysis. All strains of M. tuberculosis used were cultured in Middlebrook

7H9 media (Difco) supplemented with 10% OADC (Difco) and 0.01% Tyloxapol (Sigma) at 37°C, to an OD600 ~ 0.6 – 0.8. M. tuberculosis H37Rv was obtained from BEI Resources (NR-123), M. tuberculosis +5YǻcydAB (52) and M. tuberculosis +5YǻcydKO A317V (20) were gifts from

Dr. Digby Warner and Dr. Helena Boshoff, respectively. M. tuberculosis oxygen consumption rate

(OCR) was measured using the Seahorse XF96 Analyzer (Agilent) as previously described (16). In short, M. tuberculosis bacilli were adhered to the bottom of a XF96 cell culture microplate (Agilent) at a density of 2x106 bacilli/well using Cell-TakTM cell adhesive (Corning). Extra cellular flux analysis was carried out in unbuffered 7H9 media, at pH 7.35, containing 0.2% glucose. Basal OCR was measured for ~19 min before the automatic addition, through the drug ports of the XF96 sensory cartridge (Agilent), of either Q203 (final concentration of 0.3 μM, 100x the MIC50) or AX-35 (final concentration of 14 μM, 100x the MIC50) to the three different M. tuberculosis strains. Q203 was a

81 gift from Dr. Helena Boshoff. The deviations from basal respiration, caused by compound addition, were measured for ~35 min before the addition of the uncoupler CCCP (final concentration of 2 μM,

Sigma) to induce maximal OCR, after which OCR was measured for a final ~19 min. The points of compound addition are indicated by vertical dotted lines. OCR data points are representative of the average OCR after 3 minutes of continuous measurement, the error calculated automatically by the

Seahorse Wave Desktop 2.3.0 software (Agilent) from the OCR measurements from at least four replicate wells. OCR plots are representative of two independent experiments performed and data representation was done using GraphPad Prism 7.02.

Data Availability. Sequence data were deposited at the NCBI Sequence Read Archive under the accession number SRP142469 (https://www.ncbi.nlm.nih.gov/sra/SRP142469). RNA-seq data were deposited at the NCBI Gene Expression Omnibus under the accession number GSE113683

(https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE113683).

82 ACKNOWLEDGEMENTS

We would like to thank Dr. Digby Warner and Dr. Helena Boshoff for providing M. tuberculosis

+5YǻcydAB and M. tuberculosis +5YǻcydKO A317V strains used in the bioenergetics flux assays. We also greatly appreciate the technical help and expertise of Dr. Claudia Sala and Dr.

Charlotte Avanzi in RNA-seq and insightful discussions, and library preparation, respectively. The research was also co-funded by the South African Medical Research Council to AJCS. The research leading to these results received funding from the European Community’s Seventh Framework

Programme (MM4TB, Grant 260872).

83 REFERENCES

1. World Health Organization. 2017. Global Tuberculosis Report 2017. S.l.

2. World Health Organization, Global Tuberculosis Programme. 2016. WHO treatment guidelines

for drug-resistant tuberculosis: 2016 update.

3. Caminero JA, Sotgiu G, Zumla A, Migliori GB. 2010. Best drug treatment for multidrug-

resistant and extensively drug-resistant tuberculosis. Lancet Infect Dis 10:621–629.

4. Lechartier B, Rybniker J, Zumla A, Cole ST. 2014. Tuberculosis drug discovery in the post-

post-genomic era. EMBO Mol Med 6:158–168.

5. Manjunatha UH, Smith PW. 2015. Perspective: Challenges and opportunities in TB drug

discovery from phenotypic screening. Bioorg Med Chem 23:5087–5097.

6. Mdluli K, Kaneko T, Upton A. 2014. Tuberculosis drug discovery and emerging targets. Ann N

Y Acad Sci 1323:56–75.

7. Andries K, Verhasselt P, Guillemont J, Göhlmann HWH, Neefs J-M, Winkler H, Gestel JV,

Timmerman P, Zhu M, Lee E, Williams P, Chaffoy D de, Huitric E, Hoffner S, Cambau E,

Truffot-Pernot C, Lounis N, Jarlier V. 2005. A Diarylquinoline Drug Active on the ATP

Synthase of Mycobacterium tuberculosis. Science 307:223–227.

8. Koul A, Dendouga N, Vergauwen K, Molenberghs B, Vranckx L, Willebrords R, Ristic Z, Lill

H, Dorange I, Guillemont J, Bald D, Andries K. 2007. Diarylquinolines target subunit c of

mycobacterial ATP synthase. Nat Chem Biol 3:323.

9. Pethe K, Bifani P, Jang J, Kang S, Park S, Ahn S, Jiricek J, Jung J, Jeon HK, Cechetto J,

Christophe T, Lee H, Kempf M, Jackson M, Lenaerts AJ, Pham H, Jones V, Seo MJ, Kim YM,

Seo M, Seo JJ, Park D, Ko Y, Choi I, Kim R, Kim SY, Lim S, Yim S-A, Nam J, Kang H, Kwon

84 H, Oh C-T, Cho Y, Jang Y, Kim J, Chua A, Tan BH, Nanjundappa MB, Rao SPS, Barnes WS,

Wintjens R, Walker JR, Alonso S, Lee S, Kim J, Oh S, Oh T, Nehrbass U, Han S-J, No Z, Lee

J, Brodin P, Cho S-N, Nam K, Kim J. 2013. Discovery of Q203, a potent clinical candidate for

the treatment of tuberculosis. Nat Med 19:1157.

10. Rybniker J, Vocat A, Sala C, Busso P, Pojer F, Benjak A, Cole ST. 2015. Lansoprazole is an

antituberculous prodrug targeting cytochrome bc1. Nat Commun 6:7659.

11. THE NIX-TB TRIAL OF PRETOMANID, BEDAQUILINE AND LINEZOLID TO TREAT

XDR-TB | CROI Conference.

12. Working Group for New TB Drugs. Pipeline | Working Group for New TB Drugs.

13. Ballell L, Bates RH, Young RJ, Alvarez-Gomez D, Alvarez-Ruiz E, Barroso V, Blanco D,

Crespo B, Escribano J, González R, Lozano S, Huss S, Santos-Villarejo A, Martín-Plaza JJ,

Mendoza A, Rebollo-Lopez MJ, Remuiñan-Blanco M, Lavandera JL, Pérez-Herran E, Gamo-

Benito FJ, García-Bustos JF, Barros D, Castro JP, Cammack N. 2013. Fueling Open-Source

Drug Discovery: 177 Small-Molecule Leads against Tuberculosis. Chemmedchem 8:313–321.

14. Singh A, Jain S, Gupta S, Das T, Tyagi AK. 2003. mymA operon of Mycobacterium

tuberculosis: its regulation and importance in the cell envelope. FEMS Microbiol Lett 227:53–

63.

15. Daniel J, Deb C, Dubey VS, Sirakova TD, Abomoelak B, Morbidoni HR, Kolattukudy PE.

2004. Induction of a Novel Class of Diacylglycerol Acyltransferases and Triacylglycerol

Accumulation in Mycobacterium tuberculosis as It Goes into a Dormancy-Like State in

Culture. J Bacteriol 186:5017–5030.

85 16. Lamprecht DA, Finin PM, Rahman MA, Cumming BM, Russell SL, Jonnala SR, Adamson JH,

Steyn AJC. 2016. Turning the respiratory flexibility of Mycobacterium tuberculosis against

itself. Nat Commun 7:12393.

17. Lechartier B, Hartkoorn RC, Cole ST. 2012. In Vitro Combination Studies of Benzothiazinone

Lead Compound BTZ043 against Mycobacterium tuberculosis. Antimicrob Agents Chemother

56:5790–5793.

18. Berube BJ, Parish T. 2018. Combinations of Respiratory Chain Inhibitors Have Enhanced

Bactericidal Activity against Mycobacterium tuberculosis. Antimicrob Agents Chemother

62:e01677-17.

19. Abrahams KA, Cox JAG, Spivey VL, Loman NJ, Pallen MJ, Constantinidou C, Fernández R,

Alemparte C, Remuiñán MJ, Barros D, Ballell L, Besra GS. 2012. Identification of Novel

Imidazo[1,2-a]pyridine Inhibitors Targeting M. tuberculosis QcrB. PLoS ONE 7.

20. Arora K, Ochoa-Montaño B, Tsang PS, Blundell TL, Dawes SS, Mizrahi V, Bayliss T,

Mackenzie CJ, Cleghorn LAT, Ray PC, Wyatt PG, Uh E, Lee J, Barry CE, Boshoff HI. 2014.

Respiratory Flexibility in Response to Inhibition of Cytochrome c Oxidase in Mycobacterium

tuberculosis. Antimicrob Agents Chemother 58:6962–6965.

21. Chandrasekera NS, Berube BJ, Shetye G, Chettiar S, O’Malley T, Manning A, Flint L, Awasthi

D, Ioerger TR, Sacchettini J, Masquelin T, Hipskind PA, Odingo J, Parish T. 2017. Improved

Phenoxyalkylbenzimidazoles with Activity against Mycobacterium tuberculosis Appear to

Target QcrB. ACS Infect Dis 3:898–916.

22. Moraski GC, Seeger N, Miller PA, Oliver AG, Boshoff HI, Cho S, Mulugeta S, Anderson JR,

Franzblau SG, Miller MJ. 2016. Arrival of Imidazo[2,1-b]thiazole-5-carboxamides: Potent

Anti-tuberculosis Agents That Target QcrB. ACS Infect Dis 2:393–398.

86 23. Phummarin N, Boshoff HI, Tsang PS, Dalton J, Wiles S, Barry 3rd CE, Copp BR. 2016. SAR

and identification of 2-(quinolin-4-yloxy)acetamides as Mycobacterium tuberculosis

cytochrome bc 1 inhibitors. Medchemcomm 7:2122–2127.

24. Subtil FT, Villela AD, Abbadi BL, Rodrigues-Junior VS, Bizarro CV, Timmers LFSM, de

Souza ON, Pissinate K, Machado P, López-Gavín A, Tudó G, González-Martín J, Basso LA,

Santos DS. 2017. Activity of 2-(quinolin-4-yloxy)acetamides in mycobacterium tuberculosis

clinical isolates and identification of their molecular target by whole genome sequencing. Int J

Antimicrob Agents.

25. van der Westhuyzen R, Winks S, Wilson CR, Boyle GA, Gessner RK, Soares de Melo C,

Taylor D, de Kock C, Njoroge M, Brunschwig C, Lawrence N, Rao SPS, Sirgel F, van Helden

P, Seldon R, Moosa A, Warner DF, Arista L, Manjunatha UH, Smith PW, Street LJ, Chibale K.

2015. Pyrrolo[3,4-c]pyridine-1,3(2H)-diones: A Novel Antimycobacterial Class Targeting

Mycobacterial Respiration. J Med Chem 58:9371–9381.

26. Boshoff HIM, Myers TG, Copp BR, McNeil MR, Wilson MA, Barry CE. 2004. The

Transcriptional Responses of Mycobacterium tuberculosis to Inhibitors of Metabolism NOVEL

INSIGHTS INTO DRUG MECHANISMS OF ACTION. J Biol Chem 279:40174–40184.

27. Puustinen A, Finel M, Haltia T, Gennis RB, Wikström M. 1991. Properties of the two terminal

oxidases of Escherichia coli. Biochemistry (Mosc) 30:3936–3942.

28. Matsoso LG, Kana BD, Crellin PK, Lea-Smith DJ, Pelosi A, Powell D, Dawes SS, Rubin H,

Coppel RL, Mizrahi V. 2005. Function of the Cytochrome bc1-aa3 Branch of the Respiratory

Network in Mycobacteria and Network Adaptation Occurring in Response to Its Disruption. J

Bacteriol 187:6300–6308.

87 29. Shi L, Sohaskey CD, Kana BD, Dawes S, North RJ, Mizrahi V, Gennaro ML. 2005. Changes in

energy metabolism of Mycobacterium tuberculosis in mouse lung and under in vitro conditions

affecting aerobic respiration. Proc Natl Acad Sci 102:15629–15634.

30. Small JL, Park SW, Kana BD, Ioerger TR, Sacchettini JC, Ehrt S. 2013. Perturbation of

Cytochrome c Maturation Reveals Adaptability of the Respiratory Chain in Mycobacterium

tuberculosis. mBio 4:e00475-13.

31. Berney M, Hartman TE, Jacobs WR. 2014. A Mycobacterium tuberculosis Cytochrome bd

Oxidase Mutant Is Hypersensitive to Bedaquiline. mBio 5.

32. Kalia NP, Hasenoehrl EJ, Rahman NBA, Koh VH, Ang MLT, Sajorda DR, Hards K, Grüber G,

Alonso S, Cook GM, Berney M, Pethe K. 2017. Exploiting the synthetic lethality between

terminal respiratory oxidases to kill Mycobacterium tuberculosis and clear host infection. Proc

Natl Acad Sci 114:7426–7431.

33. Lu P, Asseri AH, Kremer M, Maaskant J, Ummels R, Lill H, Bald D. 2018. The anti-

mycobacterial activity of the cytochrome bcc inhibitor Q203 can be enhanced by small-

molecule inhibition of cytochrome bd. Sci Rep 8:2625.

34. Segal W, Bloch H. 1956. Biochemical Differentiation of Mycobacterium Tuberculosis Grown

in Vivo and in Vitro. J Bacteriol 72:132–141.

35. Daniel J, Maamar H, Deb C, Sirakova TD, Kolattukudy PE. 2011. Mycobacterium tuberculosis

Uses Host Triacylglycerol to Accumulate Lipid Droplets and Acquires a Dormancy-Like

Phenotype in Lipid-Loaded Macrophages. PLOS Pathog 7:e1002093.

88 36. Fisher MA, Plikaytis BB, Shinnick TM. 2002. Microarray Analysis of the Mycobacterium

tuberculosis Transcriptional Response to the Acidic Conditions Found in Phagosomes. J

Bacteriol 184:4025–4032.

37. Betts JC, Lukey PT, Robb LC, McAdam RA, Duncan K. 2002. Evaluation of a nutrient

starvation model of Mycobacterium tuberculosis persistence by gene and protein expression

profiling. Mol Microbiol 43:717–731.

38. Hards K, Robson JR, Berney M, Shaw L, Bald D, Koul A, Andries K, Cook GM. 2015.

Bactericidal mode of action of bedaquiline. J Antimicrob Chemother 70:2028–2037.

39. Hards K, McMillan DGG, Schurig-Briccio LA, Gennis RB, Lill H, Bald D, Cook GM. 2018.

Ionophoric effects of the antitubercular drug bedaquiline. Proc Natl Acad Sci 115:7326–7331.

40. Kang S, Kim RY, Seo MJ, Lee S, Kim YM, Seo M, Seo JJ, Ko Y, Choi I, Jang J, Nam J, Park

S, Kang H, Kim HJ, Kim J, Ahn S, Pethe K, Nam K, No Z, Kim J. 2014. Lead Optimization of

a Novel Series of Imidazo[1,2-a]pyridine Amides Leading to a Clinical Candidate (Q203) as a

Multi- and Extensively-Drug-Resistant Anti-tuberculosis Agent. J Med Chem 57:5293–5305.

41. Palomino J-C, Martin A, Camacho M, Guerra H, Swings J, Portaels F. 2002. Resazurin

Microtiter Assay Plate: Simple and Inexpensive Method for Detection of Drug Resistance in

Mycobacterium tuberculosis. Antimicrob Agents Chemother 46:2720–2722.

42. Makarov V, Neres J, Hartkoorn RC, Ryabova OB, Kazakova E, Šarkan M, Huszár S, Piton J,

Kolly GS, Vocat A, Conroy TM, Mikušová K, Cole ST. 2015. The 8-Pyrrole-Benzothiazinones

Are Noncovalent Inhibitors of DprE1 from Mycobacterium tuberculosis. Antimicrob Agents

Chemother 59:4446–4452.

89 43. Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence

data. Bioinformatics 30:2114–2120.

44. Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods

9:357–359.

45. Garrison E, Marth G. 2012. Haplotype-based variant detection from short-read sequencing.

ArXiv12073907 Q-Bio.

46. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, Miller CA, Mardis ER, Ding

L, Wilson RK. 2012. VarScan 2: Somatic mutation and copy number alteration discovery in

cancer by exome sequencing. Genome Res 22:568–576.

47. van Kessel JC, Marinelli LJ, Hatfull GF. 2008. Recombineering mycobacteria and their phages.

Nat Rev Microbiol 6:851–857.

48. Biasini M, Bienert S, Waterhouse A, Arnold K, Studer G, Schmidt T, Kiefer F, Cassarino TG,

Bertoni M, Bordoli L, Schwede T. 2014. SWISS-MODEL: modelling protein tertiary and

quaternary structure using evolutionary information. Nucleic Acids Res 42:W252–W258.

49. Shrödinger LLC. 2015. The PyMOL molecular graphics systems, version 1.8.

50. Liao Y, Smyth GK, Shi W. 2014. featureCounts: an efficient general purpose program for

assigning sequence reads to genomic features. Bioinformatics 30:923–930.

51. Love MI, Huber W, Anders S. 2014. Moderated estimation of fold change and dispersion for

RNA-seq data with DESeq2. Genome Biol 15:550.

52. Moosa A, Lamprecht DA, Arora K, Barry CE, Boshoff HIM, Ioerger TR, Steyn AJC, Mizrahi

V, Warner DF. 2017. Susceptibility of Mycobacterium tuberculosis Cytochrome bd Oxidase

90 Mutants to Compounds Targeting the Terminal Respiratory Oxidase, Cytochrome c.

Antimicrob Agents Chemother 61:e01338-17.

91 TABLES AND FIGURES

Table 1 Characterization of AX series: activity against M. tuberculosis H37Rv, cytotoxicity and selective index

Anti-mycobacterial activity against M. tuberculosis H37Rv Cytotoxicity in Selective Index in vitro ex vivo in HepG2 cells (TD50 /MIC) THP-1 TD50 (μg/ml) macrophages

Compound ID Structure MIC (μg/ml) IC50 (μg/ml) O S N O AX-35 N O 0.05 0.1 50 1000

O N O S AX-36 N O 0.1 0.3 37.5 375

O N O S AX-37 N N O 0.1 0.2 25 250

O N O AX-38 N O 0.2 0.3 37.5 188

O AX-39 S N O 0.3 1.8 75 250 N N O

Rifampicin 0.0008 0.1 75 93750

92 Table 2 Activity of AX-35 and AX-36 against selected microorganisms

MIC (μg/ml) Microorganisms AX-35 AX-36 RIF Bacillus subtilis >100 >100 0.3 Candida albicans >100 >100 1.5 Corynebacterium diphtheriae >100 >100 0.0004 Corynebacterium glutamicum >100 >100 0.004 Enterococcus faecalis >100 >100 0.6 Escherichia coli 74 100 6.7 Listeria monocytogenes >100 >100 0.8 Micrococcus luteus >100 >100 0.7 Mycobacterium avium 8.2 12.8 25 Mycobacterium bovis BCG 0.8 0.8 0.0008 Mycobacterium canettii STB-L 0.003 0.01 0.002 Mycobacterium marinum 11.5 19.6 0.5 Mycobacterium massiliense 7.7 10.4 26.9 Mycobacterium smegmatis 100 63 1.7 Mycobacterium tuberculosis 0.05 0.1 0.001 Mycobacterium ulcerans 1.6 1.6 0.01 Mycobacterium vaccae 12.1 26.9 2.5 Pseudomonas aeruginosa >100 >100 1 Pseudomonas putida >100 >100 0.2 Salmonella typhimurium >100 >100 0.7 Staphylococcus aureus >100 >100 3.8

93 FIG 1 Molecular structure of AX-35 (GW861072X)

94 FIG 2 Evidence of AX compounds targeting QcrB of M. tuberculosis. (A) Dose-response curves of isolated mutants resistant to either AX-35 or AX-36 compared to wild-type H37Rv. Three independent 7H9 liquid cultures of M. tuberculosis containing 50x or 100x MIC of AX-35 or AX-36 at passage 5 were plated on 7H10 medium to obtain single colonies. WGS results of AX-resistant mutants reveal mutations in QcrB. (B) Dose- response curves of AX-resistant, Q203-resistant, LPZS-resistant mutants to AX-35, AX-36, LPZS, Q203, and

RIF for cross-resistance studies. Data plotted are presented as mean ± SD, curves are representative of at least two independent experiments. (C) The QcrB protein of M. tuberculosis was modelled on chain A of the crystal structure of the mutant Rhodobacter sphaeroides cytochrome bc1 oxidase (PDB code: 2QJK). Clustering of mutations associated with resistance to AX, Q203 and LPZS around the quinol oxidation site of QcrB, indicated by dotted blue line. (D) Intracellular ATP levels in H37Rv and AX-resistant strains were measured in the absence and presence of BDQ, Q203, AX-35 and AX-36 at 2.5x MIC after 24 h using BacTiter-GloTM

(Promega). Data from two independent experiments are presented as mean ± SD. Statistical analysis was performed using two-way ANOVA, Tukey’s multiple comparison test (****p<0.0001).

95 FIG 3 Transcriptional response of M. tuberculosis to AX-35 treatment by RNA-seq. (A) Global transcriptomic response and involvement of different metabolic responses, based on TubercuList classification

(https://mycobrowser.epfl.ch/), after exposure of two independent cultures of M. tuberculosis H37Rv to AX-35 at 10x and 30x MIC for 4 h. (B) Heat map representing top significantly differentially regulated M. tuberculosis genes (padj 0.05). Colour scale indicates differential regulation as log2 fold-change of H37Rv with AX-35 treatment relative to H37Rv with vehicle control, DMSO. Up-regulation is indicated in red, down-regulation in blue and insignificant log2 fold-change values for the condition are in grey. Data are from two independent experiments. (C) qPCR validation of genes, cydB, lipU, and tgs4 in H37Rv and AX-resistant mutant strains treated with vehicle (DMSO) or AX-35. Data are presented as mean ± SD of two independent cultures.

Statistical analysis was performed using two-way ANOVA with Tukey’s multiple comparison test (*p<0.05,

****p<0.0001).

96 FIG 4 Bioenergetic flux profiles of M. tuberculosis with AX-35 or Q203 treatment. Oxygen consumption rate of wild-type H37Rv, H37RvǻcydAB and H37RvǻcydKO with a QcrB A317V mutation (Q203-resistance SNP) measured at basal levels, then in the presence of (A) AX-35 (final concentration of 14 μM, 100x MIC50) or (B)

Q203 (final concentration of 0.3 μM, 100x MIC50), and then at maximum capability with the depolarization of the bacterial membrane by the protonophore carbonyl cyanide m-chlorophenyl hydrazine (CCCP).

97 FIG 5 Activity of AX compounds assessed in a mouse model of acute TB. One day after low-dose aerosol M. tuberculosis infection, groups of five mice were treated with vehicle controls (TPGS 20% or TPGS 20% with

1% DMSO) or with compounds administered by oral gavage for ten days daily at the doses indicated (mg/kg).

AX compounds were prepared in TPGS 20%, Q203 in TPGS 20% with 1% DMSO and INH in ddH2O.

Bacterial burden (CFU) was determined from lung homogenates. Data from one experiment is presented as mean ± SD. Statistical analysis was performed using one-way ANOVA, Tukey’s multiple comparison test (** p<0.01, ****p<0.0001). L.o.d. indicates limit of detection.

98 SUPPLEMENTAL MATERIAL - SYNTHESIS AND CHARACTERIZATION DATA

OF AX-35 AND ANALOGS

General information. All reactions were performed under an argon atmosphere using flame-dried glassware and standard syringe/septa techniques. CH2Cl2 and THF used for reactions were distilled under argon prior to use (CH2Cl2 from CaH2 and THF from

Na/benzophenone). All other absolute solvents were purchased as anhydrous grade from Acros

(puriss.; dried over molecular sieves; H2O <0.005 %) and used without further purification.

Solvents for extractions, flash column chromatography and thin layer chromatography (TLC) were either purchased as commercial grade and distilled prior to use or purchased as HPLC grade. All other commercially available reagents were used without further purification.

Reactions were magnetically stirred and monitored by TLC performed on Merck TLC aluminum sheets (silica gel 60 F254). Spots were visualized with UV light (Ȝ = 254 nm) or through staining with Ce2(SO4)3/phosphomolybdic acid/H2SO4 (CPS), or KMnO4/K2CO3.

NaHCO3 and brine (NaCl) refer to aqueous saturated solutions. Chromatographic purification of products was performed using either Sigma-Aldrich or SiliCycle silica gel 60 for preparative column chromatography (particle size 40-63 ȝm). 1H-, 13C- and 31P-NMR spectra were recorded on a Bruker AV-400 400 MHz. Chemical shifts (į) are reported in ppm and are referenced to chloroform (į 7.26 ppm for 1H, į 77.16 ppm for 13C). All 13C-NMR spectra were measured with complete proton decoupling. Data for NMR spectra are reported as follows: s = singlet, d = doublet, t = triplet, q = quartet, quint = quintet, sept = septet, m = multiplet, br = broad signal, app. = apparent. Infrared spectra (IR) were recorded on a Jasco FT/IR-6200 instrument.

Resonance frequencies are given as wavenumbers in cm-1. High resolution mass spectra

(HRMS) were recorded on a Bruker maXis (ESI) by the ETH Zürich MS service.

99 Synthesis of AX-35 and analogs. AX-35 (originally GW861072X in Ballell et al.,

2013) was synthesized in two steps. EDC-mediated amide coupling of diethylphosphonoacetic acid (1) with 1-piperonylpiperazine (2) provided phosphonate 3 in 96% yield. Subsequent

HWE-reaction with thiophen-2-carbaldehyde (4i) afforded AX-35 in 90% yield (scheme 1).

Using intermediate 3 with the corresponding aldehyde in the HWE-reaction, four different analogs, AX-36 - AX-39, were synthesized for initial SAR investigation (scheme 1).

O O O O O HN a P P EtO N O EtO OH N EtO EtO O N O 1 2 3

R 4i-v O b R N O N O AX-35 - AX-39

S S R = iiiS AX-35, R = AX-36, R = S

iii S iv AX-37, R = S AX-38, R = N N

S S v AX-39, R = N N

Scheme 1. Reagents and conditions: (a) EDC, HOBt, Et3N, DCM, 80 °C (MW), 3 h, 96%; (b)

Aldehydes 4i-v, NaH, THF, rt, 18 h, 86-98%.

100 Diethyl (2-(4-(benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-2-oxoethyl)phosphonate (3)

O O P O N O O N O

Diethylphosphonoacetic acid (1) (434.6 μL, 2.70 mmol, 1.20 eq.), HOBt (365.3 mg, 2.70 mmol,

1.20 eq.), EDC (476.9 μL, 2.70 mmol, 1.20 eq.), Et3N (374.6 μL, 2.70 mmol, 1.20 eq.), and 1- piperonylpiperazine (2) (496.2 mg, 2.25 mmol, 1.00 eq.) were dissolved in DCM (5.0 mL). The mixture was heated in the microwave at 80 °C for 3 h, quenched with saturated NaHCO3 (10 mL) and extracted with DCM (3 x 10 mL). The combined organic layers dried over MgSO4, filtered and concentrated in vacuo. Flash chromatography (DCM/MeOH 98:2 to 95:5 to 9:1) provided phosphonate 3 (864.1 mg, 2.17 mmol, 96%) as a yellow oil.

1 TLC (SiO2; DCM/MeOH 95:5, UV): Rf = 0.16. H-NMR (400 MHz, CDCl3 į SSP

(s, 1H), 6.60 – 6.55 (m, 2H), 5.78 (s, 2H), 4.01 (dq, J = 8.2, 7.1 Hz, 4H), 3.52 – 3.39 (m, 4H),

3.27 (s, 2H), 2.91 (d, J = 22.0 Hz, 2H), 2.29 (dt, J = 31.6, 5.1 Hz, 4H), 1.19 (t, J = 7.1 Hz, 6H).

13 2 C-NMR (101 MHz, CDCl3 į SSP (d, JC-P = 7.0 Hz), 147.4, 146.5, 131.2, 121.9,

2 1 109.0, 107.6, 100.7, 62.3, 62.2 (d, JC-P = 7.2 Hz), 52.6, 52.2, 46.7, 41.7, 33.0 (d, JC-P = 132.8

3 31 Hz), 16.1 (d, JC-P = 8.7 Hz). P-NMR (162 MHz, CDCl3 į SSP P IR (thin film):

,ݝ = 2980, 2924, 2813, 1641, 1502, 1489, 1441, 1394, 1367, 1341, 1247, 1139, 1095, 1028, 966

-1 + 867, 810, 789, 720, 676, 591, 576 cm . HR-MS (ESI): Calcd for C18H28N2O6P [M+H] ,

399.1679 m/z; Found, 399.1676 m/z.

101 (E)-1-(4-(Benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-3-(thiophen-2-yl)prop-2-en-1-one

(AX-35)

O S N O N O AX-35

To a solution of 3 (753.1 mg, 1.89 mmol, 1.00 eq.) in THF (25.0 mL) were added NaH (54.4 mg, 2.27 mmol, 1.20 eq.) and thiophen-2-carbaldehyde (4i) (210.1 μL, 2.27 mmol, 1.20 eq.).

The reaction mixture was stirred at rt for 18 h, filtered through celite, then concentrated in vacuo. Flash chromatography (DCM/MeOH 98:2 to 95:5) provided AX-35 (608.5 mg, 1.71 mmol, 90%) as a yellow oil.

1 TLC (SiO2; DCM/MeOH 98:2, UV): Rf = 0.22. H-NMR (400 MHz, CDCl3): į (ppm) = 7.76

(dt, J = 15.1, 0.8 Hz, 1H), 7.25 (dt, J = 5.1, 1.0 Hz, 1H), 7.15 (dt, J = 3.5, 0.8 Hz, 1H), 6.97 (dd,

J = 5.1, 3.6 Hz, 1H), 6.83 – 6.80 (m, 1H), 6.72 – 6.67 (m, 2H), 6.63 (d, J = 15.1 Hz, 1H), 5.88

(s, 2H), 3.62 (d, J = 46.2 Hz, 4H), 3.38 (s, 2H), 2.42 – 2.36 (m, 4H). 13C-NMR (101 MHz,

CDCl3): į (ppm) = 165.1, 147.9, 146.9, 140.6, 135.6, 131.7, 130.3, 128.1, 127.2, 122.32, 116.0,

,IR (thin film): ݝ = 3069, 3005, 2895 .42.3 ,45.9 ,52.7 ,53.3 ,62.6 ,62.7 ,101.1 ,108.0 ,109.5

2810, 2772, 1638, 1597, 1501, 1488, 1439, 1416, 1367, 1335, 1298, 1282, 1269, 1238, 1205,

1145, 1114, 1096, 1037, 1000, 966, 932, 922, 864, 822, 810, 791, 703, 651, 575, 534, 518, 490

-1 + cm . HR-MS (ESI): Calcd for C19H21N2O3S [M+H] , 357.1267 m/z; Found, 357.1264 m/z.

102 (E)-1-(4-(Benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-3-(thiophen-3-yl)prop-2-en-1-one

(AX-36)

S N O N O AX-36

To a solution of 3 (1.51 g, 3.79 mmol, 1.00 eq.) in THF (50.0 mL) were added NaH (109.0 mg,

4.54 mmol, 1.20 eq.) and thiophen-3-carbaldehyde (4ii) (415.0 μL, 4.54 mmol, 1.20 eq.). The reaction mixture was stirred at rt for 18 h, filtered through celite, then concentrated in vacuo.

Flash chromatography (DCM/MeOH 98:2 to 95:5) provided AX-36 (1.30 g, 3.64 mmol, 96%) as a yellow oil.

1 TLC (SiO2; DCM/MeOH 95:5, UV): Rf = 0.17. H-NMR (400 MHz, CDCl3): į (ppm) = 8.18

(d, J = 15.3 Hz, 1H), 7.94 (dd, J = 2.8, 1.3 Hz, 1H), 7.83 – 7.78 (m, 2H), 7.37 (d, J = 1.3 Hz,

1H), 7.29 – 7.17 (m, 3H), 6.45 (s, 2H), 4.28 – 4.09 (m, 4H), 3.94 (s, 2H), 2.99 – 2.92 (m, 4H).

13 C-NMR (101 MHz, CDCl3): į (ppm) = 165.4, 147.7, 146.7, 138.2, 136.3, 131.5, 126.8, 126.7,

,IR (thin film): ݝ = 3675 .42.1 ,45.8 ,52.6 ,53.1 ,62.5 ,100.9 ,107.9 ,109.3 ,116.7 ,122.1 ,125.1

3077, 2987, 2971, 2900, 2811, 2772, 2238, 1731, 1645, 1599, 1502, 1488, 1439, 1412, 1368,

1335, 1298, 1283, 1270, 1239, 1146, 1114, 1094, 1038, 1000, 975, 908, 867, 809, 781, 727,

-1 + 645, 606, 573, 524, 425 cm . HR-MS (ESI): Calcd for C19H21N2O3S [M+H] , 357.1267 m/z;

Found, 357.1271 m/z.

103 (E)-1-(4-(Benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-3-(1O3,3O2-thiazol-4-yl)prop-2-en-

1-one (AX-37)

S N O N N O AX-37

To a solution of 3 (1.51 g, 3.79 mmol, 1.00 eq.) in THF (50.0 mL) were added NaH (109.2 mg,

4.55 mmol, 1.20 eq.) and thiazole-4-carbaldehyde (4iii) (515.0 mg, 4.55 mmol, 1.20 eq.). The reaction mixture was stirred at rt for 18 h, filtered through celite, then concentrated in vacuo.

Flash chromatography (DCM/MeOH 98:2 to 95:5) provided AX-37 (1.32 g, 3.70 mmol, 98%) as a yellow oil.

1 TLC (SiO2; DCM/MeOH 95:5, UV): Rf = 0.14. H-NMR (400 MHz, CDCl3): į (ppm) = 8.83

– 8.74 (m, 1H), 7.64 (dd, J = 14.9, 0.8 Hz, 1H), 7.40 (d, J = 2.0 Hz, 1H), 7.30 (d, J = 15.0 Hz,

1H), 6.84 (d, J = 1.3 Hz, 1H), 6.76 – 6.68 (m, 2H), 5.93 (s, 2H), 3.69 (dt, J = 28.2, 4.8 Hz, 4H),

13 3.41 (s, 2H), 2.46 – 2.40 (m, 4H). C-NMR (101 MHz, CDCl3): į (ppm) = 165.2, 153.7, 153.6,

153.2, 153.1, 147.7, 146.8, 134.2, 131.5, 122.2, 120.2, 119.8, 109.4, 107.9, 100.9, 62.6, 53.2,

,IR (thin film): ݝ = 3078, 2891, 2811, 2772, 2236, 1646, 1604, 1501, 1488 .42.2 ,45.9 ,52.6

1439, 1367, 1334, 1267, 1248, 1238, 1204, 1145, 1114, 1095, 1038, 999, 973, 909, 879, 819,

-1 + 791, 726, 645, 594, 574, 528, 486 cm . HR-MS (ESI): Calcd for C18H20N3O3S [M+H] ,

358.1220 m/z; Found, 358.1223 m/z.

104 (E)-1-(4-(Benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-3-phenylprop-2-en-1-one (AX-38)

N O N O AX-38

To a solution of 3 (1.50 g, 3.77 mmol, 1.00 eq.) in THF (50.0 mL) were added NaH (108.4 mg,

4.52 mmol, 1.20 eq.) and benzaldehyde (4iv) (459.0 μL, 4.52 mmol, 1.20 eq.). The reaction mixture was stirred at rt for 18 h, filtered through celite, then concentrated in vacuo. Flash chromatography (DCM/MeOH 98:2 to 95:5) provided AX-38 (1.24 g, 3.53 mmol, 94%) as a yellow oil.

1 TLC (SiO2; DCM/MeOH 95:5, UV): Rf = 0.26. H-NMR (400 MHz, CDCl3): į (ppm) = 7.65

(d, J = 15.4 Hz, 1H), 7.54 – 7.46 (m, 2H), 7.40 – 7.28 (m, 3H), 6.92 – 6.81 (m, 2H), 6.79 – 6.64

(m, 2H), 5.92 (s, 2H), 3.68 (d, J = 39.6 Hz, 4H), 3.42 (s, 2H), 2.53 – 2.36 (m, 4H). 13C-NMR

(101 MHz, CDCl3): į (ppm) = 165.4, 147.8, 146.8, 142.7, 135.3, 131.6, 129.6, 128.8, 127.8,

,IR (thin film): ݝ = 3060 .42.2 ,45.9 ,52.6 ,52.6 ,53.2 ,62.6 ,101.0 ,107.9 ,109.4 ,117.2 ,122.2

2810, 2772, 2236, 1646, 1603, 1500, 1488, 1438, 1366, 1334, 1301, 1238, 1205, 1145, 1115,

1096, 1038, 999, 976, 920, 863, 810, 762, 727, 705, 685, 645, 565, 530, 482 cm-1. HR-MS

+ (ESI): Calcd for C21H23N2O3 [M+H] , 351.1703 m/z; Found, 351.1707 m/z.

105 (E)-1-(4-(Benzo[d][1,3]dioxol-5-ylmethyl)piperazin-1-yl)-3-(thiazol-2-yl)prop-2-en-1-one

(AX-39)

O S N O N N O AX-39

To a solution of 3 (1.76 g, 4.42 mmol, 1.00 eq.) in THF (50.0 mL) were added NaH (127.2 mg,

5.30 mmol, 1.20 eq.) and thiazole-2-carbaldehyde (4v) (465.6 μL, 5.30 mmol, 1.20 eq.). The reaction mixture was stirred at rt for 18 h, filtered through celite, then concentrated in vacuo.

Flash chromatography (DCM/MeOH 98:2 to 95:5) provided AX-39 (1.35 g, 3.78 mmol, 86%) as an orange oil.

1 TLC (SiO2; DCM/MeOH 95:5, UV): Rf = 0.38. H-NMR (400 MHz, CDCl3): į (ppm) = 7.83

(d, J = 3.2 Hz, 1H), 7.69 (d, J = 15.1 Hz, 1H), 7.36 (d, J = 3.2 Hz, 1H), 7.26 (d, J = 15.1 Hz,

1H), 6.81 (d, J = 1.3 Hz, 1H), 6.72 – 6.67 (m, 2H), 5.90 (s, 2H), 3.65 (dt, J = 33.7, 4.9 Hz, 4H),

13 3.39 (s, 2H), 2.44 – 2.38 (m, 4H). C-NMR (101 MHz, CDCl3): į (ppm) = 164.2, 147.9, 146.9,

144.5, 133.5, 131.6, 122.3, 121.7, 121.1, 109.5, 108.1, 101.1, 62.7, 53.3, 52.6, 46.1, 42.5. IR

,thin film): ݝ = 3448, 3075, 2900, 2811, 2773, 1733, 1643, 1605, 1501, 1487, 1439, 1397, 1367)

1334, 1289, 1238, 1137, 1114, 1094, 1037, 999, 966, 932, 875, 810, 791, 768, 733, 707, 656,

-1 + 617, 590, 577, 519 cm . HR-MS (ESI): Calcd for C18H20N3O3S [M+H] , 358.1220 m/z; Found,

358.1227 m/z.

106 1H-, 13C- AND 31P-NMR SPECTRA

O O P O N O O N O

3 6.00 2.00 2.00 4.00 2.00 4.05 4.00 1.00 2.00

9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

107 147.44 146.46 131.24 121.88 109.04 107.58 100.65 62.30 62.24 62.17 52.63 52.15 46.70 41.73 33.69 32.37 162.83 162.76 16.14 16.06

O O P O N O O N O

200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

108 21.13

O O P O N O O N O

140 120 100 80 60 40 20 0 -20 -40 -60 -80 -100 -120 -140 -160 -180 -200 -220 -240 f1 (ppm)

109 O S N O N O AX-35 1.00 1.00 1.00 1.00 1.00 2.00 1.00 2.00 4.00 2.00 4.00

10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

110 147.85 146.88 140.62 135.63 131.66 130.29 128.10 127.23 122.32 115.95 109.50 108.04 101.06 62.69 62.64 53.31 52.71 45.93 42.34 165.09

O S N O N O AX-35

200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

111 O

S N O N O AX-36 1.00 1.00 2.00 1.00 3.06 2.00 4.00 2.00 4.00

10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

112 147.68 146.70 138.24 136.31 131.50 126.84 126.65 125.06 122.14 116.70 109.32 107.86 100.89 62.50 53.12 52.56 45.75 42.14 165.40

S N O N O AX-36

210 200 190 180 170 160 150 140 130 120 110 100 102030405060708090 0 f1 (ppm)

113 O

S N O N N O AX-37 1.00 1.00 1.00 1.00 1.00 2.00 2.00 4.00 2.00 4.00

10.5 10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

114 165.15 153.18 153.11 147.74 146.76 134.18 131.54 122.19 120.20 119.78 109.37 107.93 100.94 62.55 53.24 52.58 45.85 42.24 153.68 153.60

S N O N N O AX-37

200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

115 O

N O N O AX-38 1.00 2.00 3.00 2.00 2.00 2.00 4.00 2.00 4.00

9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

116 165.36 147.76 146.79 142.65 135.34 131.55 129.59 128.81 127.76 122.22 117.22 109.40 107.94 100.96 62.59 53.20 52.63 52.59 45.88 42.23

N O N O AX-38

200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

117 O S N O N N O AX-39 1.00 1.00 1.00 1.10 1.00 2.09 2.00 4.16 2.00 4.18

10.0 9.5 9.0 8.5 8.0 7.5 7.0 6.5 6.0 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 f1 (ppm)

118 147.88 146.92 144.53 133.52 131.60 122.32 121.65 121.05 109.48 108.07 101.08 62.65 53.31 52.63 46.09 42.47 164.24

O S N O N N O AX-39

200 190 180 170 160 150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0 f1 (ppm)

119 SUPPLEMENTARY METHODS

Assessment of efficacy in a chronic mouse model of TB. Female BALB/c mice (5-6 weeks old, 20g) from Charles River Laboratories were infected with a low-dose aerosol of M. tuberculosis H37Rv. Treatment was initiated 4 weeks after the infection by oral gavage 5 days a week over 4 weeks. INH was prepared at 25 mg/kg in ddH2O, Q203 at 10 mg/kg in TPGS

20% (as described in (1) and AX derivatives at 10 mg/kg in TPGS 20%. The non-treated group received the vehicle TPGS 20% alone. Compounds were ground with a pestle and mortar followed by sonication at room temperature for 30 min. Compound solutions were stored at 4oC and freshly prepared at the start of each week. The day after the final treatment, all groups were sacrificed and serial dilutions of lung and spleen homogenates were plated on 7H10 agar containing 10 μg/ml cyclohexamide and 25 μg/ml ampicillin. Experiments were approved by the Swiss Cantonal Veterinary Authority (authorisation number 3082).

Checkerboard assays for two-drug combinations containing AX-35. Interactions between AX-35 and PBTZ169, BDQ or CFM were determined using the checkerboard assay

(2, 3). M. tuberculosis H37Rv was grown to log-phase in 7H9 media and diluted to an OD600 of 0.0001. 75 μl of bacterial suspension (about 103 cells) was added per well of a 96-well plate.

PBTZ169, BDQ or CFM were two-fold serially diluted column-wise (1st compound). AX-35 was prepared in 7H9 medium starting at 8x MIC and serial dilutions were made to 0.125x. 25

μl of diluted compound at each concentration was added to a row of the 96-well plate. After incubation of the plates for 6 days at 37oC, 10 μl of 0.025% resazurin was added. The fluorescence intensity was read after 24 h incubation using an Infinite F200 Tecan plate reader.

For rows where an MIC value could be determined, the fractional inhibitory concentration index

(ȈFIC index) was calculated using the equation ȈFIC index = FIC1st compound + FICAX-35 = (MIC of 1st compound, tested in combination)/(MIC of 1st compound, alone) + (MIC of AX-35, tested

120 in combination)/(MIC of AX-35, alone). ȈFIC index 0.5 indicates synergism, 0.5 < ȈFIC index DGGLWLYLW\DQGȈFIC index > 4 antagonism.

121 REFERENCES

1. Pethe K, Bifani P, Jang J, Kang S, Park S, Ahn S, Jiricek J, Jung J, Jeon HK, Cechetto J, Christophe T, Lee H, Kempf M, Jackson M, Lenaerts AJ, Pham H, Jones V, Seo MJ, Kim YM, Seo M, Seo JJ, Park D, Ko Y, Choi I, Kim R, Kim SY, Lim S, Yim S-A, Nam J, Kang H, Kwon H, Oh C-T, Cho Y, Jang Y, Kim J, Chua A, Tan BH, Nanjundappa MB, Rao SPS, Barnes WS, Wintjens R, Walker JR, Alonso S, Lee S, Kim J, Oh S, Oh T, Nehrbass U, Han S-J, No Z, Lee J, Brodin P, Cho S-N, Nam K, Kim J. 2013. Discovery of Q203, a potent clinical candidate for the treatment of tuberculosis. Nat Med 19:1157.

2. Reddy VM, Einck L, Andries K, Nacy CA. 2010. In Vitro Interactions between New Antitubercular Drug Candidates SQ109 and TMC207. Antimicrob Agents Chemother 54:2840–2846.

3. Lechartier B, Hartkoorn RC, Cole ST. 2012. In Vitro Combination Studies of Benzothiazinone Lead Compound BTZ043 against Mycobacterium tuberculosis. Antimicrob Agents Chemother 56:5790–5793.

122 SUPPLEMENTARY TABLES

Table S1 Intrinsic clearance (Clint) and clearance category of AX derivatives in mouse and human liver microsomes

Mouse microsomes Human microsomes

Clearance Clint Clearance Clint

Category (μl/min/mg) Category (μl/min/mg) AX-35 high 119.6 medium 21.7

AX-36 high 116.8 medium 16.8

AX-37 N.D. N.D. N.D. N.D.

AX-38 high 63.8 medium 13.3

AX-39 N.D. N.D. N.D. N.D.

Carbamazepine low 0.6 low 0.4

Nifedipine high 104.6 high 121.1 N.D. = not determined

123 Table S2 Interactions of AX-35 with PBTZ169, BDQ or CFM in M. tuberculosis H37Rv. Data obtained are mean ȈFIC indices ± SD from two independent experiments.

AX-35 with PBTZ169 AX-35 with BDQ AX-35 with CFM fold MIC of DĞĂŶɇ&/ DĞĂŶɇ&/ DĞĂŶɇ&/ AX-35 SD SD SD /ŶĚĞǆ /ŶĚĞǆ /ŶĚĞǆ 0.5 1.63 0.08 1.06 0.13 0.79 0.07 0.25 1.31 0.04 1.33 0.00 0.80 0.15 0.125 1.03 0.24 1.17 0.00 0.92 0.29 0.0625 0.83 0.28 1.00 0.17 1.02 0.13

124 Table S3 List of primers used in this study

Primer name Sequence

RvQcrB_S182P_For 5’-CTGTCGGGACTCGGTCTGCGCGCGGCACTCTCGCCGATCACGCTGGGTATGCCGGTAATCGGGACCTGGC- 3’ (recombineering) RvQcrB_M342V_For 5’-ACCATTCCCGCCCCGGTCTGGGTCGCCGTGATCGTGGGCCTGGTTTTCGTCCTGCTACCCGCCTACCCAT-3’

(recombineering) YĐƌͺĨƵůůͺϬϭ& 5’-AATCCTGTGCCCTTGTCACC-3’

QcrB_full_01R 5’-AAAATGCGCCGGACTTGAAC-3’

YĐƌͺĨƵůůͺϬϮ& 5’-CCATCTTGATCCCCAGGCTC-3’

QcrB_full_02R 5’-AGAAAACTGCCACTACCCGG-3’

YĐƌͺĨƵůůͺϬϯ& 5’-TGGTTTTCGTCCTGCTACCC-3’

QcrB_full_03R 5’-GGTGATCGAGTGGCTATACG-3’

ĐǇĚͺ& 5’-GACGATGCCTACCGATTCGC-3’

cydB_R 5’-CCAGCCACGTCCAGTCTTTG-3’

ůŝƉhͺ& 5’-CAAAGGAACACAAGCAGGCG-3’

lipU_R 5’-GTCTACCTGGTTCCTCGCTG-3’

ƚŐƐϰͺ& 5’-GTCACCTTCGCCAGCATCAA-3’

tgs4_R 5’-GGTTTGGAGCTCGGTGAATG-3’

125 SUPPLEMENTARY FIGURES

FIG S1 Recombineering of mutations associated with AX resistance in chromosomal M. tuberculosis

H37Rv qcrB. (A) Two recombinant clones with strong resistance phenotype to AX compared to wild- type H37Rv or H37Rv containing co-transformed vector pYUB412. (B) Sanger sequencing across qcrB confirms insertion of desired base changes in recombinant clones resulting in missense mutations S182P and M342V. The absence of other mutations was confirmed by WGS.

126

FIG S2 Mode of action of AX-35. MBCs (representing 99.9% killed bacteria compared to D0) of AX-

35 were measured in M. tuberculosis H37Rv, H37RvǻcydAB, and H37RvǻcydAB::cydAB complemented strains by exposing mycobacteria to varying concentrations of AX-35 (fold MICs respective to each strain) over a duration of 7 days in liquid 7H9 complete media followed by plating on 7H10 solid medium. CFUs were counted after 4 weeks of incubation at 37oC.

127

FIG S3 Activity of AX compounds in a chronic mouse model of TB. Bacterial burden (CFU) was determined in the lungs (black columns) and spleens (grey columns) of five mice treated with vehicle control (TPGS 20%) or compounds treated with doses as indicated brackets (mg/kg) by oral gavage.

Day 0 indicates the start of treatment, whereas the rest of the data are obtained from Day 28 when treatment ended. Data from one experiment is presented as mean ± SD. Statistical analysis was performed using one-way ANOVA, Dunnett’s Multiple Comparison Test compared to non-treated vehicle condition (*p<0.05, ***p<0.001).

128 Chapter 3: Discovery of new plant-derived inhibitors with potent F420-dependent activity against Mycobacterium tuberculosis

Caroline S. Foo1, Tobias Brütsch2, Patrick Eisenring2, Anthony Vocat1,Andrej Benjak1, Andréanne Lupien1, Jérémie Piton1, Karl-Heinz Altmann2, Stewart T. Cole1

1Global Health Institute, École Polytechnique Fédérale de Lausanne, Switzerland

2Institute of Pharmaceutical Sciences, Department of Chemistry and Applied Biosciences, ETH

Zürich, Switzerland

Manuscript in preparation, 2018

Contributions: design and execution of experiments, data analysis, manuscript preparation

129

ABSTRACT

As Tuberculosis (TB) continues to be a global pandemic, new drugs are urgently needed against its etiological agent, Mycobacterium tuberculosis. Herein, we report on our work with a new, promising series of anti-tuberculars, PB compounds, which have been derived from naturally-occurring lapachol.

PB analogs have substantially improved activity against replicating M. tuberculosis in vitro compared to lapachol, and demonstrate activity against non-replicating M. tuberculosis SS18b as well as in infected THP-1 macrophages. The activity of PB analogs is selective and not associated with cytotoxicity. Isolation and characterization of mutants of M. tuberculosis resistant to PB compounds revealed mutations in the fgd1 gene encoding the F420-dependent glucose-6-phosphate dehydrogenase, a likely primary determinant of PB-resistance. Cross-resistance studies further identified the lack of

F420 biosynthesis in mediating resistance to PB analogs. Interestingly, a M. tuberculosis strain without a functional deazaflavin-dependent nitroreductase Ddn, which relies on reduced F420 for enzymatic activity, remained susceptible to PB compounds. This presents a different mechanism of action compared to the prodrugs Delamanid and PA-824, both of which require reduced F420 and Ddn for activation via nitro-reduction. PB compounds therefore act through a novel mechanism of action and are likely prodrugs which require bioactivation by a different reduced- F420-dependent enzyme to highly active species. Our findings highlight the potential of these lapachol-derived PB analogs as new, promising, potent anti-tuberculars with a novel F420-dependent activity and will guide future lead optimisation and target identification.

131 INTRODUCTION

Tuberculosis (TB) is one of the top ten most deadly diseases globally and the leading cause of death due to an infectious disease (1), with Mycobacterium tuberculosis (M. tuberculosis) as its etiological agent. The current front-line treatment for drug-susceptible TB (DS-TB) emerged in the 1980s, and is a six-month, directly-observed therapy (DOT) comprising isoniazid, rifampicin, pyrazinamide and ethambutol. The treatment for drug-resistant TB (DR-TB) is more complex (consisting of at least five drugs), lasts even longer (between 9 and 20 months), and is costly (between US$2000 and 5000 per individual) with low cure rates (at around 50% or less) (2). Within an infected host, M. tuberculosis is thought to exist in a spectrum of metabolic states (3). Non-replicating bacilli, in particular, persist against conventional anti-TB drugs, which are highly effective against replicating bacilli (4), and explain the lengthy TB treatment durations required to achieve cure. Adherence to prolonged treatment regimens is often poor and this contributes towards the emergence of drug-resistant M. tuberculosis strains. As part of new regimens, new anti-TB drugs should thus be potent with novel mechanisms of action to reduce treatment durations, target different bacillary sub-populations including persisters, and kill drug-resistant strains more effectively (5).

Instances of new anti-TB drugs are Bedaquiline (BDQ, Sirturo) and Delamanid (Deltyba), whose approvals in 2012 by the U.S. Food and Drug Administration (FDA) (6) and 2014 by the European

Commission (7), respectively, were important milestones in the context of contemporary TB drug development after a dry spell of new TB drugs for more than 40 years. While BDQ targets the atpE subunit of the ATP synthase (8), Delamanid has a more pleiotropic mechanism of action. Delamanid is a bicyclic nitroimidazole that shares a similar mechanism of action to PA-824 (Pretomanid), another related nitroimidazole, which is currently in Phase 3 clinical trials (Fig. 1a, b) (9). Both compounds are active against replicating and non-replicating M. tuberculosis (10, 11). Their aerobic activity has been attributed to the inhibition of mycolic acid synthesis, whereas their anaerobic activity stems from their bioactivation by a reduced F420-dependent enzyme, a deazaflavin-dependent nitroreductase

132 (Ddn), to highly active intermediates that kill M. tuberculosis non-specifically through the release of nitric oxide, which causes respiratory poisoning and cellular damage of the pathogen (10–13).

Resistance to these compounds is mediated by mutations in fbiA, fbiB, and fbiC, three genes involved in biosynthesis of the 5-deazaflavin F420 cofactor , and fgd1 whose product generates the reduced form of F420, and Ddn (11, 14, 15).

As with Delamanid and PA-824, most compounds in the global TB drug pipeline are small, synthetic, drug-like molecules. Another strategy to generate new chemical entities is to exploit natural products for their potential as anti-tuberculars. A historical example is the front-line TB drug rifampicin, which was generated from semisynthetic modifications of rifamycins, a class of naturally-occurring antibiotics isolated from Streptomyces mediterranei (18). Similar approaches have been adopted in contemporary TB drug discovery, for instance based on Streptomyces-derived spectinomycin (19), griselimycin (20), capuramycin (21), and pyridomycin (22). Analogs generated from these natural products are being optimised for improved potency against M. tuberculosis and to enhance their pharmalogical properties.

Apart from bacteria, another important source of natural medicinal compounds comes from plants.

Lapachol, a naturally-occuring naphthaquinone (Fig. 1c), is extracted from trees of the Bignoniaceae family, including those from the Tabebuia genus which are native to Central and South America (23).

Like other quinones, a wide spectrum of medicinal properties have been attributed to lapachol, including anti-cancer, anti-microbial, anti-malarial, and anti-fungal activity (24, 25). Due to its availability and simple structure which is suitable for SAR studies, we considered lapachol as an attractive starting scaffold for lead optimisation against M. tuberculosis.

In this study, we describe the characterisation of four lapachol-derived analogs, the PB compounds, which are active against M. tuberculosis in vitro and in infected macrophages. PB analogs have markedly improved activity which is selective against M. tuberculosis compared to the parent

133 compound, lapachol, and also demonstrate ex vivo activity in infected macrophages. Isolation of M. tuberculosis strains resistant to PB compounds revealed a dependence on the reduced form of the F420 cofactor for their activity.

RESULTS

Characterization of lapachol-derived PB compounds. The activity of the naturally- occurring compound lapachol against M. tuberculosis H37Rv in vitro was initially measured. Since lapachol is active against replicating M. tuberculosis H37Rv in vitro with a MIC of 13.9 μg/ml (Fig.

1, Table 1), an SAR study was initiated and several lapachol-derived analogs, named PB, were generated. All PB analogs have substantially improved activity against M. tuberculosis H37Rv compared to lapachol, with MIC values ranging from 0.4 μg/ml (Table 1) to 0.05 μg/ml, with PB-117 and PB-118 being the most active analogs. To determine whether PB compounds were also active against non-replicating M. tuberculosis, the percentage of maximum inhibition (Imax) at 10 μg/ml of these molecules was measured in vitro using the streptomycin-starved M. tuberculosis 18b strain as a latency model (26). As shown in Table 1, Imax values of these compounds range from 35% and 96%, indicating some activity of the PB family in this model of non-replicating M. tuberculosis.

To further characterise the antimicrobial activity of the PB series, two analogs, namely PB-089 and

PB-116, were tested against a panel of bacteria and fungi. Intriguingly, their activity was found to be highly specific to certain mycobacteria of the M. tuberculosis complex (MTBC), namely M. tuberculosis, M. canettii, M. bovis, and M. bovis BCG (Table 2), implying that the mechanism of action of PB compounds is related to specific properties of these strains from the MTBC complex.

Since M. tuberculosis is an intracellular pathogen, the anti-microbial activity of the compounds were measured in an ex vivo model of M. tuberculosis-infected THP-1 macrophages and IC50 values were measured to determine intracellular activity. IC50 ranged from 1.4 to 3.2 μg/ml for PB-089, PB-117, and PB-118 (Table 1), reflecting their potent ex vivo activity. However, no ex vivo activity for PB-116

134 was observed in this model. Cytotoxicity was measured in human HepG2 cells, and no to moderate cytotoxicity was observed for these compounds, as evident from the TD50 values obtained (Table 1).

This results in acceptable selectivity of all PB analogs, with selective indices (HepG2 TD50/ MIC) above 250, and PB-089 and PB-117 being the two most selective analogs (Table 1). To obtain insight into the metabolic stability of PB compounds, the intrinsic clearance (Clint) was determined in mouse and human liver microsomes. As evident from Table 3, PB analogs have only low metabolic stability in both cases.

Altogether, structural modifications to lapachol led to substantially improved anti-tubercular activity of the PB compounds, with PB-117 and PB-118 as leads of this series.

Deleterious mutations in fgd1 are associated with resistance to PB compounds.

Spontaneous resistant mutants were isolated in M. tuberculosis H37Rv on 7H10 complete solid medium containing compounds at concentrations of either 10x MIC (for PB-089) or 20x MIC (for PB-

116, PB-117, and PB-118). MIC values for isolated resistant clones were determined in 7H9 liquid medium using the resazurin microtiter assay plate (REMA) method, and PB-resistant clones were found to have a 10- to 40-fold increase in MIC compared to wild-type H37Rv. Resistance frequencies of M. tuberculosis H37Rv to PB compounds in vitro are estimated at between 1 x 10-6 and 8 x 10-7.

Whole-genome sequencing (WGS) of PB-resistant strains revealed two mutations in each strain, one occurring in fgd1 and the other in a pps gene, resulting in missense mutations, premature polypeptide termination or frameshifts (Table 4). In M. tuberculosis, fgd1 encodes the F420-dependent glucose-6- phosphate dehydrogenase, Fgd1, which catalyses the oxidation of glucose-6-phosphate to 6- phosphogluconolactone with the concomitant reduction of the F420 cofactor while pps genes are involved in the biosynthesis of the cell wall lipid phthiocerol dimycocerosate (PDIM), which is involved in the virulence of the bacillus (27). All identified mutations associated with resistance to PB are functionally deleterious for their respective gene functions, including Ser54Leu in fgd1 as predicted

135 by the Provean web server (28, 29), which is in line with the non-essentiality of both fgd1 and pps genes for in vitro growth of M. tuberculosis (30).

Cross-resistance studies of PB compounds and PB-resistant strains. To investigate if all analogs in the PB series share the same mechanism of action, the susceptibility of PB-resistant mutants to all PB compounds were determined. From dose-response curves of the two PB-resistant strains to all PB analogs, an increase in MIC by 20- to 100-fold compared to that in wild-type H37Rv was observed (Fig. 2), indicating that all analogs share the same mechanism of action.

Since deleterious mutations in Fgd1 and in the enzymes involved in the F420 biosynthesis pathway are also known to mediate resistance to Delamanid and PA-824, cross-resistance studies to Delamanid and

PA-824 were also conducted to gain further insight into the mechanism of action of PB compounds.

As expected, PB-resistant mutants also demonstrate complete resistance to Delamanid and PA-824

(Fig. 2), further validating the absence of a functional Fgd1 in these strains. The activity of PB compounds was additionally determined in two Delamanid/PA-824 resistant mutants, namely an FbiC mutant strain, which has a transposon insertion in fbiC, and a Ddn mutant harbouring a transposon in ddn, with defective FbiC and Ddn gene products, respectively. Interestingly, while the fbic mutant strain is highly resistant to all PB analogs (Fig. 2), which would imply that the F420 cofactor is necessary for PB activity as with Delamanid and PA-824, the ddn mutant strain has no or only very low resistance to PB compounds (Fig. 2), indicating that the mechanism of action of this family does not largely depend on Ddn, unlike Delamanid and PA-824.

Altogether, the cross-resistance studies indicate that all PB analogs share the same mechanism of action against M. tuberculosis in vitro and confirm the presence of fgd1 mutations in PB-resistant strains. Furthermore, the mechanism of action of PB compounds differs from that of Delamanid and

PA-824, despite their activation requiring the F420 cofactor, as the nitroreductase Ddn is not involved.

136 DISCUSSION

In this study, we describe the characterisation of four lapachol-derived PB analogs with substantially improved activity against replicating M. tuberculosis in vitro compared to the parent scaffold. The potent micromolar activity of these compounds is selective against M. tuberculosis and not due to cytotoxicity, and limited to mycobacterial species of the MTBC tested. PB compounds demonstrate activity against non-replicating M. tuberculosis in the SS18b model, and most analogs have activity in

M. tuberculosis-infected macrophages. A potential liability of these compounds that will need addressing during further optimisation, is their low metabolic stability, as observed in mouse and human microsomes, which could reduce their efficacy against M. tuberculosis in vivo.

Similar to PA-824-resistance, mutations in fgd1 and pps were identified in M. tuberculosis strains spontaneously resistant to PB compounds (11, 12). Although not validated in this study, it is unlikely that pps genes are primary determinants of PB resistance. pps genes are required for the biosynthesis of PDIM, in which mutations have often been observed in in vitro experiments (31). This is probably attributed to the large coding capacity necessary for the production of PDIM and its in vitro redundancy as a virulence factor (27, 31, 32). Furthermore, it was demonstrated that PDIM mutations were not associated with PA-824 resistance as validated by genetic complementation (12).

It is interesting to note that PDIM synthesis in mycobacteria relies on the reduced form of the 5- deazaflavin electron carrier, F420-H2 (33). F420 production is uncommon and restricted to certain archaea and bacteria (34, 35), including universally in mycobacteria (36). M. tuberculosis is predicted to carry at least 33 F420-dependent enzymes (28, 34), including F420-dependent glucose 6-phosphate dehydrogenase Fgd1 and deazaflavin-dependent nitroreductase Ddn, both of which have been experimentally validated (37, 38). Fgd1 is an unusual glucose-6-phosphate dehydrogenase in that it catalyses the oxidation of glucose-6-phosphate to 6-phosphogluconolactone via F420 instead of NADP

137 (37, 39) and is a main enzyme for the generation of reduced F420 (40, 41), while the proposed physiological role of Ddn is that of a F420-H2-dependent quinone reductase (38).

Fgd1, together with Ddn and the F420 biosynthetic enyzmes FbiA, FbiB, and FbiC, are necessary components for the bioactivation of the prodrugs Delamanid and PA-824. Mutations in any of these enzymes render M. tuberculosis resistant to these compounds (11, 14, 15). In the case of PB analogs, only non-functional Fgd1 and FbiC enzymes result in PB resistance, indicating that activity of these compounds relies on F420 and its reduced form, without the involvement of Ddn. This is consistent with the absence of available nitro groups on PB compounds for nitro-reduction (Table 1). It is highly likely that PB compounds are also prodrugs, activated by another hitherto unidentified F420-H2- dependent enzyme, which is common to M. tuberculosis, M. bovis and M. canettii, members of the

MTBC, and absent in other mycobacteria, to highly reactive species and causing pleiotropic damage to the bacillus without acting on a specific target, in a similar manner to Delamanid and PA-824 (Fig.

3).

Since nitroreductases are uninvolved in the activation of PB compounds, 26 F420-dependent enzymes identified up until now could act as potential activators, among which two are Fgd enzymes and the others are monooxygenases and oxidoreductases. From comparative mycobacterial genomics, the involvement of Rv3618 and Rv2074 can be ruled out as although these genes have been deleted from the M. bovis genome, this bacterium remains susceptible to PB analogs. An additional 14 genes are unlikely to be involved as well, as they are present in M. smegmatis, against which PB compounds are inactive. This leaves 8 possible candidates that could serve as the activator and investigating them further should help us to understand the activation mechanism and to identify the active form of the

PB compounds.

A pertinent issue with the use of quinones as therapeutics is their associated cytotoxicity (25). As PB compounds require F420-H2-dependent activation, it is unlikely that they retain broad-spectrum activity

138 as F420 and Fgd1 production is restricted to certain bacteria (39), thus likely reducing perturbations of the host microbiota. Furthermore, since mammalian enzymes are incapable of F420 biosynthesis, cross- activation of these compounds leading to unspecific toxic side effects on the host could be avoided as well, thus mitigating the risk of quinone cytotoxicity.

Since a variety of mutations in seemingly inessential genes, fgd1, fbiA, fbiB, and fbiC mediate in vitro resistance to PB, the in vitro resistance frequency of M. tuberculosis to PB compounds is high and estimated at 10-6 to 10-7. This is comparable to the ranges reported for Delamanid, PA-824, and INH at 10-5 to 10-6, 10-5 to 10-7, and 10-6 respectively (12, 14, 15, 42). Notably, as mutations in ddn are not associated with resistance to PB compounds, PB analogs have the potential to be backup molecules for Delamanid and PA-824 for strains of M. tuberculosis resistant to these compounds due to mutations in ddn.

The clinical relevance of these mutations, however, remain to be determined as Delamanid and PA-

824 advance in Phase 3 clinical trials (9). In a case of a clinical isolate, mutations in fbiA and fgd1 have been associated with phenotypic resistance to Delamanid (43). Since F420-H2 possibly contributes to low-redox potential reactions for the anaerobic survival of M. tuberculosis (44), and fgd1 and F420-H2 have roles in protecting the bacillus against nitrosative and oxidative stress (40, 41), it is tempting to speculate that mutations in fgd1 and/or F420 biosynthetic enzymes would occur at a higher fitness cost for certain bacillary populations, and therefore the in vivo emergence of resistance might be reduced or the transmission of such resistant strains be impaired.

In conclusion, it is evident from our SAR data that structural changes to the natural product, lapachol, can markedly improve activity against M. tuberculosis and generate a promising new family of anti- tuberculars. It is likely that PB compounds act through a novel mechanism of action against M. tuberculosis. Altogether, considering the profiling performed in this study, the PB analogs merit

139 further optimisation and characterisation studies, and ongoing work is also being undertaken to identify their activator and the active species.

140 TABLES AND FIGURES

Table 1 Characterization of PB series: activity against M. tuberculosis H37Rv and SS18b; cytotoxicity and

selective index

Anti-mycobacterial Anti-mycobacterial activity against M. activity against M. Cytotoxicity tuberculosis H37Rv tuberculosis SS18b Selective in HepG2 Index in vitro ex vivo in in vitro cells THP-1 macrophages MIC TD50 Compound ID Structure IC (μg/ml) I (%) (TD /MIC) (μg/ml) 50 max (μg/ml) 50

N PB-089 N 0.2 ~3.2 35 >100 500 O

O O

N PB-116 0.1 no activity 47 28 280 N O O O

O N

N PB-117 O O 0.05 ~2.5 96 21 420 O

N PB-118 O O 0.05 1.4 96 13 260 O

PB-001 N O O 0.4 N.D. 75 >100 250 (TB-189) O

O OH Lapachol 13.9 N.D. N.D. N.D. N.D.

Rifampicin 0.001 0.1 71 50 50000

141 Table 2 Activity of PB-089 and PB-116 against selected microorganisms

MIC (μg/ml) Microorganisms PB-089 PB-116 RIF

Bacillus subtilis >100 >100 0.3 Candida albicans >100 >100 1.5 Corynebacterium diphtheriae >100 >100 0.0004 Corynebacterium glutamicum >100 >100 0.004 Enterococcus faecalis >100 >100 0.6 Escherichia coli >100 >100 6.7 Listeria monocytogenes >100 >100 0.8 Micrococcus luteus >100 >100 0.7 Mycobacterium abscessus >100 >100 22.5 Mycobacterium avium >100 >100 25 Mycobacterium bolletii >100 >100 47 Mycobacterium bovis AF2122/97 0.2 0.7 0.0008 Mycobacterium bovis BCG 0.3 0.4 0.0008 Mycobacerium canettii STB-L 0.1 0.3 0.01 Mycobacterium marinum >100 >100 0.5 Mycobacterium massiliense >100 >100 26.9 Mycobacterium smegmatis >100 >100 1.7 Mycobacterium tuberculosis 0.2 0.1 0.001 Mycobacterium ulcerans >100 >100 0.01 Mycobacterium vaccae >100 >100 2.5 Pseudomonas aeruginosa >100 >100 1 Pseudomonas putida >100 >100 0.2 Salmonella typhimurium >100 >100 0.7 Staphylococcus aureus >100 >100 3.8

142 Table 3 Intrinsic clearance (Clint) and clearance category of AX derivatives in mouse and human liver microsomes

Mouse microsomes Human microsomes

Clearance Cl Clearance Cl int int Category (μl/min/mg) Category (μl/min/mg)

PB-089 high 462.9 high 63.8

PB-116 high 105.7 high 115.9

PB-117 high N.D. high 471.1

PB-118 high 64.9 high 220.1

Carbamazepine low 2.9 low 0.4

Nifedipine high 116.2 high 116.2

N.D. = could not be determined

143 Table 4 Whole-genome sequencing (WGS) of M. tuberculosis H37Rv PB-resistant strains

Selected on 7H10 Codon Amino acid Compound plates + compound Clone Gene change change (fold MIC) fgd1 161C>T Ser54Leu PB 089 10x A4 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 089 10x A6 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 117 20x 1 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 117 20x 5 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 116 20x B1 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 116 20x B4 ppsA 2871dupC Glu958fs fgd1 161C>T Ser54Leu PB 118 20x C4 ppsA 2871dupC Glu958fs fgd1 55G>T Glu19* PB 118 20x C3 ppsD 4743dupG Gln1582fs

144 FIG 1 Chemical structures of (A) Delamanid, (B) PA-824 (Pretomanid), and (C) lapachol, from which PB analogs are derived

145 FIG 2 Cross-resistance studies of PB compounds. (A) Dose-response curves of M. tuberculosis wild-type

H37Rv, FbiC mutant, Ddn mutant, and two PB-resistant mutant strains to PB compounds, Delamanid, PA-824, and RIF. Data are presented as technical duplicates of mean ± SD; graphs are representative of two independent experiments. (B) Overview of the fold increase in MIC for each compound in each strain as a ratio to the MIC obtained for M. tuberculosis H37Rv.

146 FIG 3 Proposed mechanism of F420-dependent action of PB compounds. The biosynthesis of the 5-deazaflavin hydride carrier F420 involves FbiA, FbiB, and FbiC, a process restricted to certain species of bacteria, including mycobacteria. Fgd1 reduces F420 with the oxidation of glucose-6-phosphate to 6-phosphogluconolactone.

Reduced F420, F420-H2, is utilised by an unidentified F420-dependent enzyme (not Ddn) to activate PB compounds to highly active intermediates which are detrimental for M. tuberculosis, either killing or inhibiting the growth of the bacillus. Adapted from (15).

147 MATERIALS AND METHODS

Drugs used in this study. Lapachol and RIF were from Sigma Aldrich.

Culture conditions of M. tuberculosis strains, other bacteria and eukaryotic cell lines.

Corynebacterium diphtheriae, Corynebacterium glutamicum, Escherichia coli, Micrococcus luteus,

Pseudomonas putida, Salmonella typhimurium,and Staphylococcus aureus were grown in LB broth.

Enterococcus faecalis, Listeria monocytogenes, and Pseudomonas aeruginosa were grown in brain heart infusion (BHI) broth. HepG2 cells were grown in DMEM (Gibco) media supplemented with

o 10% fetal bovine serum at 37 C with 5% CO2. THP-1 macrophages were grown in RPMI medium

o supplemented with 10% fetal bovine serum and 1 mM sodium pyruvate at 37 C with 5% CO2.

Determination of MICs. MICs were determined using the resazurin reduction microplate assay (REMA) as previously described (45). Strains were grown to log-phase (OD600 0.4 to 0.8) and diluted to an OD600 of 0.0001. Two-fold serial dilutions of each test compound were prepared in 96- well plates containing 100 μl of bacteria per well (3 x 103 cells per well). Plates were incubated either at 30oC or 37oC as required with appropriate incubation times (e.g. for M. tuberculosis at 37oC, 6 days).

(SS18b) was performed as described above for REMA assays using an SS18b culture maintained in

7H9 medium without streptomycin for 2 weeks and a final OD600 of 0.1 as previously reported (46).

148 Cytotoxicity for HepG2 cells. Human HepG2 cells (4,000 cells/well) were incubated for 3

Assessment of drug activity ex vivo in THP-1 macrophages. 1 x 105 THP-1 human monocytic cells/well were seeded in 96-well plates and incubated with 4 nM phorbol-12-myristate-13- acetate (PMA) overnight to stimulate macrophage differentiation. Differentiated macrophages were infected with M. tuberculosis H37Rv grown to log phase (OD600 0.4 to 0.8) at MOI 5. Extracellular

o bacteria were removed after 3-4 h incubation at 37 C with 5% CO2 by removing the RPMI medium and washing with PBS. Compounds to be tested were prepared in separate 96-well plates by two-fold serial dilutions in a final volume of 100 μl RPMI which was then transferred to the plates of infected

o THP-1 macrophages. Plates were sealed and incubated for 48 h at 37 C with 5% CO2. 10 μlof

PrestoBlue® (ThermoFischer Scientific) was added and plates were incubated for up to 1 h at 37oC with 5% CO2 before the fluorescence intensity (excitation/emission: 560/590 nm) was measured using

Infinite F200 Tecan plate reader. Dose-response curves were plotted and IC50 values obtained using a non-linear regression fit equation (log[inhibitor] vs response, variable slope) in GraphPad Prism.

Microsomal stability studies. Metabolic stability of the compounds was measured based on intrinsic clearance (Clint) in mouse and human liver microsomes as previously described (47). Final compound concentration in the mixture of microsomes and NADPH-regeneration system was 2 μg/ml, and a mixture without NADPH-regeneration was also prepared for each compound as a control of the stability of the compound with time. Carbamazepine and nifedipine at 2 μg/ml were used as low and high intrinsic clearance controls, respectively.

Isolation and characterization of PB-resistant mutants. PB-resistant mutants of M. tuberculosis H37Rv were isolated on 7H10 complete solid agar plates containing compound

149 concentrations at 10x MIC for PB-089 and 20x MIC for PB-116, PB-117, and PB-118, with 2.4 x 10E8 bacteria exposed per condition. REMA assays were performed on isolated colonies after re-streaking on 7H10 complete solid medium. Genomic DNA extraction was performed using the QiaAMP UCP

Pathogen Minikit (Qiagen) as per manufacturer’s instructions. Whole-genome sequencing was performed using Illumina technology with sequencing libraries prepared using the KAPA HyperPrep kit (Roche) and sequenced on Illumina HiSeq 2500 instrument. All raw reads were adapter- and quality-trimmed with Trimmomatic v0.33 (48) and mapped onto the M. tuberculosis H37Rv reference genome (RefSeq NC_000962.3) using Bowtie2 v2.2.5 (49). The bamleftalign program from the

FreeBayes package v0.9.20-18 (50) was used to left-align indels. Reads with mapping quality below

8 and duplicate reads were omitted.

Variant analysis. Variant calling was done using VarScan v2.3.9 (51) using the following cut- offs: minimum overall coverage of ten non-duplicated reads, minimum of five non-duplicated reads supporting the SNP, base quality score >15, and a SNP frequency above 30%. The rather low thresholds, especially the SNP frequency, were deliberately chosen to avoid missing potential variants in alignment-difficult regions, or in case of mixed population. All putative variants unique to the mutant strains were manually checked by inspecting the alignments.

150 ACKNOWLEDGEMENTS

We would like to thank Prof. Cliff Barry and Dr. Neeraj Dhar for providing M. tuberculosis Ddn and

FbiC transposon mutant strains, and Dr. Jaroslav Roh for the synthesis of PA-824 used in this study.

We also appreciate the technical expertise of Dr. Charlotte Avanzi for the library preparation. The research leading to these results received funding from the European Community’s Seventh

Framework Programme (MM4TB, Grant 260872).

151 REFERENCES

1. 2017. The top 10 causes of death. WHO.

2. World Health Organization. 2017. Global Tuberculosis Report 2017. S.l.

3. Mitchison DA. 1979. Basic Mechanisms of Chemotherapy. Chest 76:771–780.

4. McKinney JD. 2000. In vivo veritas: The search for TB drug targets goes live. Nat Med

6:1330–1333.

5. Zumla A, Nahid P, Cole ST. 2013. Advances in the development of new tuberculosis drugs and

treatment regimens. Nat Rev Drug Discov 12:388–404.

6. Sirturo (bedaquiline) product insert. Silver Spring, MD: Food and Drug Administration.

7. European Medicines Agency - Find medicine - Deltyba.

8. Koul A, Dendouga N, Vergauwen K, Molenberghs B, Vranckx L, Willebrords R, Ristic Z, Lill

H, Dorange I, Guillemont J, Bald D, Andries K. 2007. Diarylquinolines target subunit c of

mycobacterial ATP synthase. Nat Chem Biol 3:323.

9. Working Group for New TB Drugs |.

10. Matsumoto M, Hashizume H, Tomishige T, Kawasaki M, Tsubouchi H, Sasaki H, Shimokawa

Y, Komatsu M. 2006. OPC-67683, a Nitro-Dihydro-Imidazooxazole Derivative with Promising

Action against Tuberculosis In Vitro and In Mice. PLOS Med 3:e466.

11. Stover CK, Warrener P, VanDevanter DR, Sherman DR, Arain TM, Langhorne MH, Anderson

SW, Towell JA, Yuan Y, McMurray DN, Kreiswirth BN, Barry CE, Baker WR. 2000. A small-

molecule nitroimidazopyran drug candidate for the treatment of tuberculosis. Nature 405:962–

966.

152 12. Manjunatha UH, Boshoff H, Dowd CS, Zhang L, Albert TJ, Norton JE, Daniels L, Dick T,

Pang SS, Barry CE. 2006. Identification of a nitroimidazo-oxazine-specific protein involved in

PA-824 resistance in Mycobacterium tuberculosis. Proc Natl Acad Sci U S A 103:431–436.

13. Manjunatha U, Boshoff HI, Barry CE. 2009. The mechanism of action of PA-824. Commun

Integr Biol 2:215–218.

14. Haver HL, Chua A, Ghode P, Lakshminarayana SB, Singhal A, Mathema B, Wintjens R, Bifani

P. 2015. Mutations in Genes for the F420 Biosynthetic Pathway and a Nitroreductase Enzyme

Are the Primary Resistance Determinants in Spontaneous In Vitro-Selected PA-824-Resistant

Mutants of Mycobacterium tuberculosis. Antimicrob Agents Chemother 59:5316–5323.

15. Fujiwara M, Kawasaki M, Hariguchi N, Liu Y, Matsumoto M. 2018. Mechanisms of resistance

to delamanid, a drug for Mycobacterium tuberculosis. Tuberculosis 108:186–194.

16. Schatz A, Bugle E, Waksman SA. 1944. Streptomycin, a Substance Exhibiting Antibiotic

Proc Soc Exp Biol Med .†כ.Activity Against Gram-Positive and Gram-Negative Bacteria

55:66–69.

17. Hoagland DT, Liu J, Lee RB, Lee RE. 2016. New agents for the treatment of drug-resistant

Mycobacterium tuberculosis. Adv Drug Deliv Rev 102:55–72.

18. Sensi P. 1983. History of the development of rifampin. Rev Infect Dis 5 Suppl 3:S402-406.

19. Lee RE, Hurdle JG, Liu J, Bruhn DF, Matt T, Scherman MS, Vaddady PK, Zheng Z, Qi J,

Akbergenov R, Das S, Madhura DB, Rathi C, Trivedi A, Villellas C, Lee RB, Rakesh,

Waidyarachchi SL, Sun D, McNeil MR, Ainsa JA, Boshoff HI, Gonzalez-Juarrero M, Meibohm

B, Böttger EC, Lenaerts AJ. 2014. Spectinamides: a new class of semisynthetic antituberculosis

agents that overcome native drug efflux. Nat Med 20:152–158.

153 20. Kling A, Lukat P, Almeida DV, Bauer A, Fontaine E, Sordello S, Zaburannyi N, Herrmann J,

Wenzel SC, König C, Ammerman NC, Barrio MB, Borchers K, Bordon-Pallier F, Brönstrup M,

Courtemanche G, Gerlitz M, Geslin M, Hammann P, Heinz DW, Hoffmann H, Klieber S,

Kohlmann M, Kurz M, Lair C, Matter H, Nuermberger E, Tyagi S, Fraisse L, Grosset JH,

Lagrange S, Müller R. 2015. Targeting DnaN for tuberculosis therapy using novel

griselimycins. Science 348:1106–1112.

21. Koga T, Fukuoka T, Doi N, Harasaki T, Inoue H, Hotoda H, Kakuta M, Muramatsu Y,

Yamamura N, Hoshi M, Hirota T. 2004. Activity of capuramycin analogues against

Mycobacterium tuberculosis, Mycobacterium avium and Mycobacterium intracellularein vitro

and in vivo. J Antimicrob Chemother 54:755–760.

22. Hartkoorn RC, Sala C, Neres J, Pojer F, Magnet S, Mukherjee R, Uplekar S, Boy-Röttger S,

Altmann K-H, Cole ST. 2012. Towards a new tuberculosis drug: pyridomycin – nature’s

isoniazid. EMBO Mol Med 4:1032–1042.

23. Gottlieb O R, Mors WB. 1980. Potential Utilization of Brazilian Wood Extractives. J Agric

Food Chem 28:188–196.

24. Hussain H, Krohn K, Ahmad VU, Miana GA, Green IR. 2007. Lapachol: an overview. Arkivoc

2:145–171.

25. O’Brien PJ. 1991. Molecular Mechanisms of Quinone Toxicity. Chem-Biol Interactions.

26. Sala C, Dhar N, Hartkoorn RC, Zhang M, Ha YH, Schneider P, Cole ST. 2010. Simple Model

for Testing Drugs against Nonreplicating Mycobacterium tuberculosis. Antimicrob Agents

Chemother 54:4150–4158.

154 27. Cox JS, Chen B, McNeil M, Jr WRJ. 1999. Complex lipid determines tissue-specific replication

of Mycobacterium tuberculosis in mice. Nature 402:79–83.

28. Choi Y, Sims GE, Murphy S, Miller JR, Chan AP. 2012. Predicting the Functional Effect of

Amino Acid Substitutions and Indels. PLoS ONE 7:e46688.

29. Choi Y, Chan AP. 2015. PROVEAN web server: a tool to predict the functional effect of amino

acid substitutions and indels. Bioinformatics 31:2745–2747.

30. Kapopoulou A, Lew JM, Cole ST. 2011. The MycoBrowser portal: A comprehensive and

manually annotated resource for mycobacterial genomes. Tuberculosis 91:8–13.

31. Domenech P, Reed MB. 2009. Rapid and spontaneous loss of phthiocerol dimycocerosate

(PDIM) from Mycobacterium tuberculosis grown in vitro: implications for virulence studies.

Microbiol Read Engl 155:3532–3543.

32. Camacho LR, Ensergueix D, Perez E, Gicquel B, Guilhot C. Identification of a virulence gene

cluster of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis. Mol

Microbiol 34:257–267.

33. Purwantini E, Daniels L, Mukhopadhyay B. 2016. F420H2 Is Required for Phthiocerol

Dimycocerosate Synthesis in Mycobacteria. J Bacteriol 198:2020–2028.

34. Eirich LD, Vogels GD, Wolfe RS. 1978. Proposed Structure for Coenzyme F420 from

Methanobacterium. American Chemical Society 17.

35. Daniels L, Bakhiet N, Harmon K. 1985. Widespread Distribution of a 5-deazaflavin Cofactor in

Actinomyces and Related Bacteria. Syst Appl Microbiol 6:12–17.

155 36. Selengut JD, Haft DH. 2010. Unexpected Abundance of Coenzyme F420-Dependent Enzymes

in Mycobacterium tuberculosis and Other Actinobacteria. J Bacteriol 192:5788–5798.

37. Purwantini E, Daniels L. 1996. Purification of a novel coenzyme F420-dependent glucose-6-

phosphate dehydrogenase from Mycobacterium smegmatis. J Bacteriol 178:2861–2866.

38. Gurumurthy M, Rao M, Mukherjee T, Rao SPS, Boshoff HI, Dick T, Barry CE, Manjunatha

UH. 2013. A novel F420-dependent anti-oxidant mechanism protects Mycobacterium

tuberculosis against oxidative stress and bactericidal agents. Mol Microbiol 87:744–755.

39. Purwantini E, Gillis TP, Daniels L. 1997. Presence of F420-dependent glucose-6-phosphate

dehydrogenase in Mycobacterium and Nocardia species, but absence from Streptomyces and

Corynebacterium species and methanogenic Archaea. FEMS Microbiol Lett 146:129–134.

40. Hasan MR, Rahman M, Jaques S, Purwantini E, Daniels L. 2010. Glucose 6-Phosphate

Accumulation in Mycobacteria IMPLICATIONS FOR A NOVEL F420-DEPENDENT ANTI-

OXIDANT DEFENSE SYSTEM. J Biol Chem 285:19135–19144.

41. Purwantini E, Mukhopadhyay B. 2009. Conversion of NO2 to NO by reduced coenzyme F420

protects mycobacteria from nitrosative damage. Proc Natl Acad Sci 106:6333–6338.

42. Johnson R, Streicher EM, Louw GE, Warren RM, van Helden PD, Victor TC. 2006. Drug

resistance in Mycobacterium tuberculosis. Curr Issues Mol Biol 8:97–111.

43. Bloemberg GV, Keller PM, Stucki D, Trauner A, Borrell S, Latshang T, Coscolla M, Rothe T,

Hömke R, Ritter C, Feldmann J, Schulthess B, Gagneux S, Böttger EC. 2015. Acquired

Resistance to Bedaquiline and Delamanid in Therapy for Tuberculosis. N Engl J Med

373:1986–1988.

156 44. Boshoff HIM, Barry 3rd CE. 2005. Tuberculosis — metabolism and respiration in the absence

of growth. Nat Rev Microbiol 3:70–80.

45. Palomino J-C, Martin A, Camacho M, Guerra H, Swings J, Portaels F. 2002. Resazurin

Microtiter Assay Plate: Simple and Inexpensive Method for Detection of Drug Resistance in

Mycobacterium tuberculosis. Antimicrob Agents Chemother 46:2720–2722.

46. Zhang M, Sala C, Hartkoorn RC, Dhar N, Mendoza-Losana A, Cole ST. 2012. Streptomycin-

Starved Mycobacterium tuberculosis 18b, a Drug Discovery Tool for Latent Tuberculosis.

Antimicrob Agents Chemother 56:5782–5789.

47. Makarov V, Neres J, Hartkoorn RC, Ryabova OB, Kazakova E, Šarkan M, Huszár S, Piton J,

Kolly GS, Vocat A, Conroy TM, Mikušová K, Cole ST. 2015. The 8-Pyrrole-Benzothiazinones

Are Noncovalent Inhibitors of DprE1 from Mycobacterium tuberculosis. Antimicrob Agents

Chemother 59:4446–4452.

48. Bolger AM, Lohse M, Usadel B. 2014. Trimmomatic: a flexible trimmer for Illumina sequence

data. Bioinformatics 30:2114–2120.

49. Langmead B, Salzberg SL. 2012. Fast gapped-read alignment with Bowtie 2. Nat Methods

9:357–359.

50. Garrison E, Marth G. 2012. Haplotype-based variant detection from short-read sequencing.

ArXiv12073907 Q-Bio.

51. Koboldt DC, Zhang Q, Larson DE, Shen D, McLellan MD, Lin L, Miller CA, Mardis ER, Ding

L, Wilson RK. 2012. VarScan 2: Somatic mutation and copy number alteration discovery in

cancer by exome sequencing. Genome Res 22:568–576.

157

Chapter 4: Characterization of DprE1-mediated Benzothiazinone Resistance in Mycobacterium tuberculosis

Caroline Shi-Yan Foo1, Benoit Lechartier1,2, Gaëlle S. Kolly1, Stefanie Boy-Röttger1, João Neres1,3, Jan Rybniker1,4, Andréanne Lupien1, Claudia Sala1, Jérémie Piton1, Stewart T. Cole1

1Global Health Institute, École Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland 2Centre hospitalier universitaire vaudois (CHUV), CH-1011 Lausanne, Switzerland 3UCB Biopharma, Chemin du Foriest, 1420 Braine L’Alleud, Belgium 41st Department of Internal Medicine, University of Cologne, D-50937 Cologne, Germany

Antimicrobial Agents and Chemotherapy, 2016, doi:10.1128/AAC.01523-16

Contributions: design and execution of experiments, data analysis, manuscript preparation

159

crossmark

Characterization of DprE1-Mediated Benzothiazinone Resistance in Mycobacterium tuberculosis

Caroline Shi-Yan Foo,a Benoit Lechartier,a,b Gaëlle S. Kolly,a Stefanie Boy-Röttger,a João Neres,a,c Jan Rybniker,a,d Andréanne Lupien,a Claudia Sala,a Jérémie Piton,a Stewart T. Colea Global Health Institute, École Polytechnique Fédérale de Lausanne, Lausanne, Switzerlanda; Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerlandb; UCB Biopharma, Braine L’Alleud, Belgiumc; 1st Department of Internal Medicine, University of Cologne, Cologne, Germanyd

Benzothiazinones (BTZs) are a class of compounds found to be extremely potent against both drug-susceptible and drug-resistant Mycobacterium tuberculosis strains. The potency of BTZs is explained by their specificity for their target decaprenylphosphoryl-D-ribose oxidase (DprE1), in particular by covalent binding of the activated form of the compound to the critical cysteine 387 residue of the enzyme. To probe the role of C387, we used promiscuous site-directed mutagenesis to introduce other codons at this position into dprE1 of M. tuberculosis. The resultant viable BTZ-resistant mutants were characterized in vitro, ex vivo, and biochemically to gain insight into the effects of these mutations on DprE1 function and on M. tuberculosis. Five different mutations (C387G, C387A, C387S, C387N, and C387T) conferred various levels of resistance to BTZ and exhibited different phe- notypes. The C387G and C387N mutations resulted in a lower growth rate of the mycobacterium on solid medium, which could be attributed to the significant decrease in the catalytic efficiency of the DprE1 enzyme. All five mutations rendered the mycobacterium less cytotoxic to macrophages. Finally, differences in the potencies of covalent and noncovalent DprE1 inhibitors in the presence of C387 mutations were revealed by enzymatic assays. As expected from the mechanism of action, the covalent inhibitor PBTZ169 only partially inhibited the mutant DprE1 enzymes compared to the near-complete inhibition with a noncovalent DprE1 inhibitor, Ty38c. This study emphasizes the importance of the C387 residue for DprE1 activity and for the killing action of covalent inhibitors such as BTZs and other recently identified nitroaromatic inhibitors.

ycobacterium tuberculosis is the etiological agent of tubercu- and serine, respectively. These mutations confer natural resistance Mlosis (TB), an infectious disease which is a leading cause of to BTZ (2). Spontaneous mutants resistant to BTZ that were death worldwide and poses a major threat to global health. The raised in Mycobacterium smegmatis and M. tuberculosis revealed World Health Organization estimates that in 2014, 9.6 million that glycine or serine substitutions at C387 increased the MIC by people contracted TB, and 1.5 million people died (1). In addition, at least 1,000-fold (2). The clinical importance of the C387 residue the emergence and worldwide spread of multidrug-resistant TB of DprE1 was confirmed as well when 240 M. tuberculosis clinical (MDR-TB) and extensively drug-resistant TB (XDR-TB) are isolates were tested, since all these isolates were found to be BTZ alarming. With MDR-TB strains being resistant to the frontline sensitive and had the conserved cysteine codon. drugs isoniazid and rifampin and XDR-TB strains being resistant The vulnerability of DprE1 lies in its essentiality in mycobac- to frontline and additionally second-line drugs, there is an urgent teria and its localization in the cell wall (11), accounting for the need for new drugs for TB. fact that DprE1 has been identified as the target of several struc- 1,3-Benzothiazin-4-ones (BTZs) were discovered in 2009, with turally distinct compounds in recent drug screens. These com- the lead compound BTZ043 having high potency (MIC of 1 ng/␮l) pounds can be classified as covalent or noncovalent DprE1 inhib- against M. tuberculosis strain H37Rv (2) and demonstrating effi- itors. Covalent inhibitors such as BTZ, the nitroquinoxaline cacy against MDR and XDR clinical isolates (3). Piperazine-con- VI-9376 (12), and the nitroimidazole 377790 (13) are nitroaro- taining BTZ (PBTZ) derivatives were then designed with im- matic compounds possessing the necessary nitro group required proved pharmacological properties (4), and the optimized lead for covalent adduct formation at C387 on DprE1. Noncovalent compound PBTZ169 is currently in clinical trials (5). inhibitors such as TCA1 (14), 1,4-azaindoles (15), pyrazolopyri- Genetic analysis of resistant mutants and enzymology have dones (16), 4-aminoquinolone piperidine amides (17), and Ty38c identified the target of BTZs as decaprenylphosphoryl-␤-D-ribose oxidase (DprE1), an essential flavoenzyme in M. tuberculosis involved in cell wall synthesis (2). DprE1 acts in concert with DprE2 Received 14 July 2016 Returned for modification 22 July 2016 to catalyze the epimerization of decaprenyl-phosphoribose (DPR) Accepted 7 August 2016 Accepted manuscript posted online to decaprenyl-phospho-D-arabinofuranose (DPA), which is the 15 August 2016 sole precursor for the synthesis of arabinogalactan and lipoarabi- Citation Foo CS-Y, Lechartier B, Kolly GS, Boy-Röttger S, Neres J, Rybniker J, Lupien nomannan (LAM) in the mycobacterium cell wall (6). A, Sala C, Piton J, Cole ST. 2016. Characterization of DprE1-mediated benzothiazinone resistance in Mycobacterium tuberculosis. Antimicrob Agents BTZ behaves as a suicide substrate for the reduced form of Chemother 60:6451–6459. doi:10.1128/AAC.01523-16. DprE1 by undergoing nitroreduction to form a nitroso derivative, Address correspondence to Stewart T. Cole, stewart.cole@epfl.ch. which specifically forms a covalent adduct with C387 in the DprE1 Supplemental material for this article may be found at http://dx.doi.org/10.1128 active site (7–10). The C387 residue of DprE1 is highly conserved /AAC.01523-16. in orthologous enzymes in actinobacteria, except in Mycobacte- Copyright © 2016, American Society for Microbiology. All Rights Reserved. rium avium and M. aurum, where cysteine is replaced by alanine

November 2016 Volume 60 Number 11 Antimicrobial Agents and Chemotherapy aac.asm.org 6451 161 Foo et al.

(18) block enzyme activity by forming hydrophobic, electrostatic, Fitness assessment of dprE1 mutant strains in liquid culture. All

and van der Waals interactions with particular residues in the strains were diluted to an initial OD600 of 0.05, and OD600 measurements DprE1 active site. were taken every 24 h to monitor growth over a period of 2 weeks. The Given the pivotal role played by the C387 residue of DprE1 in generation time for each strain was calculated by using the equation G ϭ ϫ the efficacy of nitroaromatic compounds, the aim of this study was t/[3.3 log10(b/B)], where G is the generation time, t is the time interval of two measurements in the exponential phase, b is the final OD , and B to identify mutations at C387 that are tolerated and confer resis- 600 is the initial OD . Two independent cultures for each strain were used tance to (P)BTZ in order to understand the underlying mecha- 600 for growth rate measurements. nisms of resistance involved as well as the overall influence of these Expression and purification of wild-type and mutant M. tuberculo- mutations on the DprE1 enzyme and on the pathogen M. tuber- sis DprE1. M. tuberculosis dprE1 was cloned into plasmid pET28a, and the culosis. recombinant protein was coexpressed in E. coli BL21(DE3) along with the MATERIALS AND METHODS M. tuberculosis GroEL2 (Rv0440) and E. coli GroES chaperones in a modified version of the pGro7 plasmid (TaKaRa Bio Inc.). DprE1 was purified Bacterial strains, culture conditions, and chemicals. M. tuberculosis as described previously (4) to obtain pure protein with bound flavin ad- 2 H37Rv, M. smegmatis mc 155, and merodiploid strains were grown at enine dinucleotide (FAD). BTZ-resistant C387G, C387S, C387A, C387T, 37°C in Middlebrook 7H9 broth (Difco) supplemented with 0.2% glyc- and C387N mutants were obtained by site-directed mutagenesis with the erol, 0.05% Tween 80, and 10% albumin-dextrose-catalase (ADC) or on pET28a-dprE1 plasmid by using the Stratagene QuikChange II site-di- Middlebrook 7H10 agar (Difco) supplemented with 0.2% glycerol and rected mutagenesis kit. Expression and purification of the mutant pro- 10% oleic acid-albumin-dextrose-catalase (OADC). For cloning proce- teins were carried out as described above for the wild-type (WT) protein. dures, One Shot TOP10 chemically competent Escherichia coli cells (In- Protein concentrations were determined by using the bicinchoninic acid vitrogen) were grown in Luria-Bertani (LB) broth or on LB agar contain- ␮ ␮ (BCA) assay (Pierce BCA protein assay kit; Thermo Scientific). ing kanamycin (50 g/ml) or hygromycin (200 g/ml). All chemicals DprE1 enzymatic activity assays. Enzyme activities of wild-type and were purchased from Sigma-Aldrich unless otherwise stated. mutant DprE1 proteins were determined in a two-step coupled assay (9). Generation of randomly mutated dprE1 in merodiploid M. tubercu- Reactions were carried out in black 96-well half-area plates (catalog num- losis strains. The dprE1 gene under the control of its natural promoter, ber 3686; Corning) in a final volume of 25 ␮l per well. The reaction located upstream of Rv3789, was amplified together with Rv3789 by using mixture consisted of the DprE1 protein (protein concentrations were primers rv3790-fwd and rv3790-rev and cloned in the pCR-Blunt II- adapted to obtain similar fluorescence signals [1.5 ␮M for the WT, 1.5 ␮M TOPO vector (Invitrogen). The resulting plasmid was used to generate for the C387S mutant, 7.5 ␮M for C387G, 3 ␮M for C387A, 10.5 ␮M for random mutations in the TGC codon encoding Cys387. Site-directed mu- C387T, and 10.5 ␮M for C387N]), FAD (1 ␮M), horseradish peroxidase tagenesis was carried out by using the Stratagene QuikChange II site- (HRP) (0.2 ␮M), Amplex Red (50 ␮M) (Life Technologies), and farnesyl- directed mutagenesis kit with primers containing random bases at the site phosphoryl-␤-D-ribofuranose (FPR) (0 ␮M, 0.2 ␮M, 0.4 ␮M, 0.6 ␮M, or of interest. The mutated fragments were ligated into the pND255 vector, 0.8 ␮M) in assay buffer (50 mM glycyl glycine [pH 8.0], 200 mM potas- kindly provided by N. Dhar, École Polytechnique Fédérale de Lausanne sium glutamate, 0.002% Brij 35). A standard curve was obtained with a (EPFL), Lausanne, Switzerland, which harbors a hygromycin resistance serial dilution of resorufin sodium salt. FPR was used to start the reaction, cassette. Six pools of randomly mutated plasmids were obtained (pBLX 1 to pBLX 6), and each pool was screened independently. The resulting and the conversion of Amplex Red to resorufin was immediately moni- integrative vectors were then transformed and integrated at the L5-attB tored by fluorescence measurement (excitation/emission wavelength of site of M. smegmatis mc2155 and M. tuberculosis H37Rv. Transformants 560/590 nm) in the kinetic mode on a Tecan M200 instrument at 30°C. To were selected on 7H10 agar plates with or without BTZ043 at 400 ng/ml. A determine 50% inhibitory concentrations (IC50s), the reaction mixture specific primer was used to repeat the site-directed mutagenesis to obtain consisting of DprE1, FAD, HRP, and Amplex Red was first incubated with the DprE1C387G mutant, which was selected on BTZ043-containing me- the test compound (with 2-fold serial dilution starting with dimethyl sul- dium. All primers are listed in Table S1 in the supplemental material. foxide [DMSO] [1% final DMSO concentration]) for 10 min at 30°C ␮ Determination of MICs. MICs were determined by using the resaz- before the addition of FPR (0.3 M). The background fluorescence inten- urin reduction microplate assay (REMA) as previously described (19). M. sity from the reaction mixture without FPR was subtracted from the val- tuberculosis strains were grown in 7H9 medium to log phase (optical den- ues for all reactions. Fluorescence units were converted to resorufin concentrations by using a standard curve. Reaction rates at each FPR sity at 600 nm [OD600] of 0.4 to 0.8) and diluted to an OD600 of 0.0001. One hundred microliters of the bacterial suspension (3 ϫ 103 cells) was concentration or compound concentration were determined and fitted to pipetted into wells of a 96-well plate. Compounds (BTZ043, PBTZ169, either the Hill equation for non-Michaelis-Menten kinetics to obtain Ty38c, moxifloxacin, and rifampin) were added to the first column, and steady-state kinetic constants or a log[inhibitor]-versus-normalized re- subsequently, 2-fold serial dilutions were made. After 6 days of incubation sponse (with 100% activity of each enzyme being defined under steady- at 37°C, 10 ␮l of 0.025% (wt/vol) resazurin was added to each well. The state conditions in the absence of the inhibitor and 0% activity being fluorescence intensity was read after 24 h of incubation by using an Infi- defined as the full inhibition of WT DprE1 in the presence of 40 ␮M the nite F200 Tecan plate reader, and MIC values were determined by non- inhibitor) to obtain IC50s by using GraphPad Prism. linear fitting of the data to the Gompertz equation (35) using GraphPad Infection of THP-1 macrophages. Human monocytic THP-1 cells Prism. were grown in RPMI medium supplemented with 10% FBS. A total of 2 ϫ Site-directed mutagenesis of dprE1 at C387 in M. tuberculosis 104 cells/well of a 96-well plate in 50 ␮l RPMI medium were differentiated H37Rv. Generation of point mutations in M. tuberculosis H37Rv was by using phorbol-12-myristate-13-acetate (PMA) (final concentration of done by a recombineering method (20, 21). H37Rv/pJV53 was grown to 4 nM). Plates were sealed with gas-permeable sealing films and incubated ␮ log phase (OD600 of 0.5) in 7H9 medium containing 25 g/ml of kana- at 37°C with 5% CO2. Cells were infected the following day. RPMI me- mycin before being induced with 0.2% acetamide overnight. Competent dium containing PMA was removed from the wells, and cells were washed cells were transformed with 100 ng of 70-bp single-stranded oligonucle- and incubated with RPMI medium alone. H37Rv, dprE1 mutant strains, ⌬ otides (leading and lagging strands) (see Table S1 in the supplemental and H37Rv RD1 were grown to log phase (OD600 of between 0.4 and ϫ 8 material) containing the desired mutations, and transformants were se- 0.8), washed in 7H9 medium, and resuspended to an OD600 of1(3 10 lected on 7H10 agar plates either with or without 400 ng/ml BTZ043. bacteria/ml). THP-1 cells were infected at a multiplicity of infection Single-nucleotide polymorphisms (SNPs) in resistant colonies were con- (MOI) of 5 in 50 ␮l RPMI medium and incubated for 3 days before the firmed by colony PCR. addition of 5 ␮l PrestoBlue cell viability reagent (Life Technologies). After

6452 aac.asm.org Antimicrobial Agents and Chemotherapy November 2016 Volume 60 Number 11 162 Mycobacterial DprE1 and BTZ Resistance

TABLE 1 DprE1 C387 mutations conferring resistance to BTZ043 in M. tuberculosis Source or organism in which mutation Codon(s) identified by Residue or mutation was previously identified No. of isolatesb BTZ043 MIC (␮g/ml) MXFc MIC (␮g/ml) sequencing C387 (WT) Wild type 0.002 0.031 TGC C387G M. tuberculosis, M. smegmatisa 3 12.5 0.063 GGA, GGC C387S M. tuberculosis, M. smegmatis, M. auruma 28 Ͼ100 0.031 TCA, TCC, TCG, TCT, AGC C387A M. aviuma 5 Ͼ100 0.031 GCC, GCG C387T This study 18 Ͼ100 0.031 ACA, ACC, ACT C387N This study 7 6.25 0.031 AAT a See reference 2. b Colonies were isolated on plates containing 400 ng/ml of BTZ043. c MXF, moxifloxacin.

1 h of incubation at room temperature (RT), fluorescence was measured levels (Table 1). Out of 9 possible codons that can be derived from by using a Tecan M200 instrument (excitation/emission wavelength of single mutations at C387, which are most likely to occur clinically, 560/590 nm). only codon TCT coding for C387S appeared in the screen. The Structural studies of WT and mutant DprE1 proteins. DprE1 mu- C387G mutation resulted in slower-growing colonies on solid me- tants were modeled based on the M. tuberculosis DprE1 structure (PDB dium, although this effect was not observed in liquid culture (see accession number 4NCR)(9), using the “Position Scan” function in the Fig. S1 and S2 in the supplemental material). The finding that the FolDX plug-in (22) implemented in YASARA View molecular graphics software (23). Illustrations were made by using PyMOL (24), and the observed phenotype was not due to additional spontaneous mu- prediction of the functional effect of amino acid substitutions was com- tations was confirmed by growing M. tuberculosis cells trans- puted by using the Provean Web server (25, 26). formed with a plasmid (pBLG) (see Table S2 in the supplemental material) harboring the C387G substitution, suggesting that the RESULTS growth defect of M. tuberculosis on solid medium was attributed to Five DprE1 mutations confer resistance to BTZ043 in M. tuber- the glycine mutation. Although the same strategy was adopted to culosis. In an initial screen to determine which mutations at the select for BTZ-resistant mutants in M. smegmatis expressing the C387 residue of DprE1 confer resistance to BTZ043, the gene with M. tuberculosis dprE1 locus, this was unsuccessful (data not a randomly mutated codon at position 387 was expressed under shown). the control of its natural promoter in M. tuberculosis H37Rv. Since To further exclude any effects of having two copies of dprE1 previous reports described BTZ-resistant substitutions at C387 as due to the partially diploid strain used in the initial screen, site- being dominant (2), a partially diploid strain was used. The pro- directed mutagenesis was performed directly on chromosomal moter of the Rv3789-dprE1-dprE2 operon (27) was exploited for dprE1 of H37Rv at the C387 residue. C387S, C387A, and C387T the construction of partially diploid strains carrying random mu- mutations were successfully introduced into dprE1 without addi- tations at the C387 residue of DprE1. In doing so, potential arti- tional undesired mutations, and normal-sized colonies were se- facts derived from any unbalanced expression of the wild-type and lected on BTZ043-containing plates (Table 2). The introduction mutant alleles were avoided. of C387G and C387N substitutions gave rise to two different col- After selection on 400 ng/ml BTZ043, transformants were ony sizes, both of which occurred in the absence of BTZ043. Small screened by PCR to identify mutations associated with resistance, colonies were found to harbor solely the desired mutation (either and their MICs for BTZ043 were determined. Various mutations C387G or C387N), whereas normal-sized colonies were found to were identified in colonies from BTZ043-containing plates, of have either a C387G V388I double mutation or C387T instead of which 28 were serine (C387S), 18 were threonine (C387T), 5 were the desired C387N mutation (Table 2). Resistance of these strains alanine (C387A), 3 were glycine (C387G), and 7 were asparagine with dprE1 C387 mutations to BTZ043 was confirmed by REMA (C387N) substitutions. Random mutagenesis proved successful, (Table 2). as several different codons for each mutation were found, and Fitness assessment of dprE1 mutant strains. The growth of BTZ043 MIC values of the colonies established their resistance H37Rv and dprE1 mutant strains was monitored in liquid 7H9

TABLE 2 Characterization of different H37Rv dprE1 mutantsa Presence of BTZ in plate from which Mutation(s) identiﬁed Mutation Codon Colony size colonies were picked by sequencing BTZ043 MIC (␮g/ml) RIF MIC (␮g/ml) C387 TGC (WT) 0.0008 0.0016 C387S TCC Normal BTZ043 C387S Ն10 0.0016 C387A GCC Normal BTZ043 C387A Ͼ10 0.0016 C387T AAT Normal BTZ043 C387T Ͼ10 0.0016 C387G GGC Small C387G Undet. Undet. Normal C387G V388I 5 0.0008 C387N ACC Small C387N 1.25 0.0008 Normal C387T* Ͼ10 0.0016 a RIF, rifampin; Undet., undetermined. * indicates the C387T mutation with an ACT codon.

November 2016 Volume 60 Number 11 Antimicrobial Agents and Chemotherapy aac.asm.org 6453 163 Foo et al.

FIG 1 Fitness of H37Rv and dprE1 mutant strains in 7H9 medium. (A) Growth curves of H37Rv and dprE1 mutant strains obtained by measuring the OD600 of bacterial cultures at 24-h intervals over a period of 14 days, with an initial OD600 of 0.05. (B) The generation time for each strain was calculated by using the ϭ ϫ equation G t/[3.3 log10(b/B)], where G is the generation time; t is the time interval between two measurements within the exponential phase, here 0 to 48 Ϯ h; b is the final OD600; and B is the initial OD600. Data from two independent experiments are presented as means standard deviations. Statistical analysis was performed by using one-way analysis of variance with Dunnett’s posttest (**, P Ͻ 0.01; ***, P Ͻ 0.001). C387T* indicates a mutation containing the ACT codon, compared to C387T, which has the AAT codon instead. “ns” indicates no significance. medium in the absence of BTZ043 over 2 weeks. Growth curves Steady-state enzymatic activity of DprE1 mutants. To inves- did not reflect any striking differences in fitness between H37Rv tigate whether DprE1 C387 mutations could affect the proper and mutant strains during the exponential or the stationary phase functioning of the enzyme, the steady-state enzymatic activity of (Fig. 1A). The generation time for each strain was calculated in the purified WT and mutant DprE1 proteins was measured by using a exponential phase of growth. For certain mutants (C387S, C387A, two-step coupled enzymatic assay. For each protein, reaction rates and C387T), their generation times significantly differ from that at four substrate concentrations were fitted to the Hill equation of H37Rv (Fig. 1B), although these differences appear to be too for non-Michaelis-Menten kinetics to obtain parameters of sub- small to be reflected in the growth curves of the mutant strains. strate binding affinity (K0.5), turnover number (kcat), and catalytic Notably, the growth defect of the C387G and C387N mutants on efficiency (kcat/K0.5). None of the mutations at C387 of DprE1 solid medium was not observed in liquid medium. affected the binding affinity of the substrate for the enzyme, as Effect of DprE1 C387 mutations on M. tuberculosis cytotox- seen from their K0.5 values, which were similar to that of the WT icity. The influence of C387 mutations on the cytotoxicity of the enzyme (Fig. 3A). C387G and C387N mutations significantly de- mycobacterium was examined ex vivo by infecting THP-1 macro- creased the turnover rate of the enzyme (Fig. 3B), and conse- phages with the H37Rv and H37Rv dprE1 C387 mutant strains. quently, these two mutations reduced the overall catalytic effi- As a control, an attenuated strain with reduced virulence, ciency of DprE1 by about 4-fold (Fig. 3C). ⌬ H37Rv RD1 (28), was used, and this showed decreased cytotox- Effect of DprE1 C387 mutations on the potency of DprE1 icity, with macrophage viability being around 60% of that of non- inhibitors. IC50 values of covalent (PBTZ169) and noncovalent infected macrophages. About 40% of macrophages remained via- (Ty38c) DprE1 inhibitors (Fig. 4A) were determined for WT and ble after infection with each of the five dprE1 mutants, compared DprE1 mutant proteins. As BTZ043 showed behavior similar to to about 25% of macrophages remaining viable after being in- that of PBTZ169 in this assay (data not shown), and the MIC of fected with H37Rv (Fig. 2), indicating that the C387 mutations PBTZ169 was 3-fold lower than that of BTZ043 in M. tuberculosis, decrease the cytotoxicity of M. tuberculosis. PBTZ169 was used in this study. Unlike the wild-type enzyme, whose activity is completely inhibited at high PBTZ169 concentrations, mutant enzymes are only partially inhibited at the highest PBTZ169 concentration (40 ␮M), apparently reaching a pla- teau at 30 to 40% activity (Fig. 4B). In contrast, this concentration of the noncovalent inhibitor Ty38c was sufficient to inhibit almost all activity of both WT and mutant enzymes (Fig. 4B and C). This finding confirms the importance of the C387 residue of DprE1 in the mode of inhibition of the covalent inhibitor PBTZ169 and indicates that the replacement of the Cys residue hardly affects the binding of noncovalent inhibitors to DprE1.

This can be further seen from the IC50 values of PBTZ169 and Ty38c obtained for WT and mutant enzymes (Fig. 4C). All mutant

enzymes showed increased PBTZ169 IC50 values compared to that for the WT, thus accounting for the BTZ resistance of the mutants. FIG 2 DprE1 C387 mutations affect the cytotoxicity of M. tuberculosis. THP-1 Ty38c IC50 values for the C387S and C387A mutants were identi- cells were infected with H37Rv, H37Rv⌬RD1, and H37Rv dprE1 C387 mutant cal to that for the WT, while the C387G, C387T, and C387N mu- strains at an MOI of 5 or left untreated (not infected [N.I.]). Macrophage tants had slightly higher values than that for the WT (Fig. 4C), viability was measured at 3 days postinfection. Data from two independent indicating that substitutions at C387 have a more modest effect on experiments were normalized to data under noninfected conditions and are presented as means Ϯ standard deviations. Statistical analysis was performed Ty38c activity. PBTZ169 and Ty38c MIC values were also deter- by using one-way analysis of variance with Dunnett’s posttest (***, P Ͻ 0.001). mined for H37Rv and dprE1 mutant strains (Table 3). C387 mu-

6454 aac.asm.org Antimicrobial Agents and Chemotherapy November 2016 Volume 60 Number 11 164 Mycobacterial DprE1 and BTZ Resistance

FIG 3 Steady-state enzymatic activity of WT and DprE1 mutants. Shown are substrate binding affinity (K0.5) (A), turnover number (kcat) (B), and catalytic efficiency (kcat/K0.5) (C) steady-state parameters obtained for the WT and DprE1 mutants by fitting enzyme activity at various FPR concentrations (0.2 to 0.8 ϭ ϫ h ϩ h h mM) to the equation Y Vmax X /(Kprime X ), where Y is enzyme activity, X is the substrate concentration, Kprime equals (K0.5) , and h is the Hill coefficient, by using GraphPad Prism. Data from at least three independent experiments are presented as means Ϯ standard deviations. Statistical analysis was performed by using one-way analysis of variance with Dunnett’s posttest (*, P Ͻ 0.05; **, P Ͻ 0.01).

tations cause an increase in the PBTZ169 MIC compared to will lead to nonfunctional proteins. More precisely, only substitu- H37Rv, with strains harboring C387S, C387A, and C387T mutations by small and/or polar residues such as asparagine, aspartate, tions having higher levels of resistance (MIC of Ͼ1 ␮g/ml) and serine, threonine, alanine, and glycine would lead to functional those with C387G and C387N mutations having lower levels of enzymes, whereas amino acids with larger side chains (arginine, resistance (MIC of 0.5 to 0.6 ␮g/ml) to PBTZ169. In the case of tryptophan, phenylalanine, glutamate, glutamine, tyrosine, me- Ty38c, mutant strains had MIC values ranging from 0.1 to 0.8 thionine, lysine, and histidine) would block access to the sub- ␮g/ml, and the trend was similar to that seen with PBTZ169. strate, resulting in nonfunctional enzymes. Furthermore, expo- Taken together, these data demonstrate that C387 mutations sure of the side chain to the solvent excludes the possibility of a decrease the potency of covalently binding DprE1 inhibitors such hydrophobic residue occupying this position (leucine, isoleucine, as PBTZ169 to a much greater extent than noncovalent DprE1 and valine), and the proline substitution is not favored in beta inhibitors, represented here by Ty38c. As the MICs of lower-resis- sheet structures, as it cannot complete the hydrogen-bonding net- tance mutants could be measured, one of the mutants (C387N) work. was used to investigate the effect of a C387 mutation on the inter- To confirm these observations, a prediction of the functional action between BTZ and bedaquiline (BDQ), as synergism be- effects of amino acid substitutions was computed by using the tween these two classes of compounds in vitro (29) and in vivo (4) Provean Web server (25, 26). Interestingly, only alanine, serine, was previously described. Using a checkerboard assay, the partial and threonine were found to be nondeleterious. Altogether, these synergism of PBTZ169 and BDQ observed in H37Rv at 0.5ϫ MIC results are in agreement with the data obtained from the screen. was not significantly altered by the DprE1 C387N mutation (see The superposition of crystal structures of WT DprE1 and the Fig. S3 in the supplemental material). C387A, C387S, C387G, C387T, and C387N mutants in complex Modeling the effect of C387 mutations on the active site of with PBTZ169 or Ty38c revealed that these five substitutions do DprE1. To gain further insight into the mechanism of resistance not block the access of both inhibitors to the binding pocket (see of DprE1 mutants to BTZ and to predict the effect of each muta- Fig. S4 in the supplemental material). As expected, the five mu- tion on the functionality of DprE1, substitution at C387 with ev- tants are not able to form the critical covalent adduct with ery amino acid was modeled in silico based on the structure of M. PBTZ169. In these cases, PBTZ169 is stabilized only unspecifically tuberculosis DprE1. by van der Waals interactions in the binding pocket and behaves as Modeling showed that all the substitutions are tolerated in an inefficient noncovalent inhibitor, resulting in a dramatic re- terms of steric hindrance. Indeed, amino acid 387 is localized on a duction of PBTZ169 potency. On the other hand, C387 is not beta sheet where the side chain could easily be accommodated directly implicated in the binding of the noncovalent inhibitor since it is exposed to the solvent toward the substrate binding Ty38c. Although C387 mutations do not influence the main in- pocket (Fig. 5A). Substitutions at this position will lead to a mod- teraction with Ty38c, they could still affect the environment in the ification of the shape and volume of the substrate binding pocket substrate binding pocket, leading to a modest impairment of the (Fig. 5B). It is evident from these models that some substitutions inhibitor activity.

November 2016 Volume 60 Number 11 Antimicrobial Agents and Chemotherapy aac.asm.org 6455 165 Foo et al.

FIG 4 Effects of DprE1 C387 mutations on potency of DprE1 inhibitor activity. (A) Structures of the DprE1 inhibitors PBTZ169 and Ty38c. PBTZ169 inhibits DprE1 via a covalent bond between the reduced form of its nitro group and the C387 residue of DprE1, whereas Ty38c, which lacks the nitro group, is a noncovalent inhibitor of DprE1. (B) Curves of enzyme activity with increasing inhibitor (PBTZ169 or Ty38c) concentrations for WT and mutant enzymes.A total of 40 ␮M the inhibitor was used for the highest concentration, with subsequent 2-fold serial dilutions. Enzymatic activities at each inhibitor concentration

were normalized to steady-state enzymatic activity in the absence of any inhibitor. (C) PBTZ169 and Ty38c IC50 values and maximum inhibition for WT and mutant enzymes. IC50 values were obtained by ﬁtting the curves in panel B to the log[inhibitor]-versus-normalized response by using GraphPad Prism. Maximum percent inhibition of the WT or mutant enzyme was determined with 40 ␮M the inhibitor. Data from at least two independent experiments are presented as means Ϯ standard deviations.

DISCUSSION the inhibitors, PBTZ169 is a highly potent covalent inhibitor of DprE1 plays an essential role in the DPA pathway for cell wall DprE1 and is currently in phase I trials. We have studied the crit- synthesis in M. tuberculosis. Having been identified as the target of ical C387 residue of DprE1 required for the covalent interaction at least five structurally distinct inhibitors in recent years, DprE1 is with PBTZ169 and with other nitroaromatic inhibitors by identi- an extremely vulnerable target of the pathogen (30, 31). Among fying and characterizing substitutions at this residue that are tolerated and confer resistance to BTZs. The data obtained confirmed the dominant nature of the mu- TABLE 3 PBTZ169 and Ty38c MICs for H37Rv and H37Rv dprE1 C387 tations introduced (2) and identified C387S, C387A, C387T, mutant strains C387G, and C387N as DprE1 mutations viable for the bacteria PBTZ169 Ty38c RIFa while conferring various levels of resistance to BTZ043 and ␮ ␮ ␮ Strain MIC ( g/ml) MIC ( g/ml) MIC ( g/ml) PBTZ169. The C387S, C387A, and C387T mutations confer high H37Rv 0.0001 0.1 0.001 levels of resistance of H37Rv to BTZs, while C387G and C387N C387S Ͼ1 0.5 0.001 confer intermediate levels of resistance. As previously reported, C387G 0.5 0.04 0.0005 C387S and C387A substitutions are naturally occurring mutations C387A Ͼ1 0.1 0.001 C387T Ͼ1 0.8 0.001 found in M. aurum and M. avium, and C387G and C387S were C387N 0.6 0.1 0.001 found in colonies spontaneously resistant to BTZ043 (2). C387N NTB1b Ͼ1 0.7 0.001 and C387T are novel mutations identified here from this screen. a RIF, rifampin. To our knowledge, these mutations have hitherto remained b BTZ-resistant strain with a C387S mutation (control). unidentified in DprE1 in mutants spontaneously resistant to

6456 aac.asm.org Antimicrobial Agents and Chemotherapy November 2016 Volume 60 Number 11 166 Mycobacterial DprE1 and BTZ Resistance

FIG 5 Closeup views of the substrate binding pocket of M. tuberculosis WT DprE1 compared to mutant models. (A) Substrate binding pocket in the crystal structure of M. tuberculosis DprE1 (PDB accession number 4NCR). FAD is represented in blue, and cysteine 387 is presented in red. The solvent-accessible surface is represented by the white surface. (B) Superposition of crystal structures of M. tuberculosis WT DprE1 and 19 substitution models at residue 387 (represented by sticks in red). The side chains of some amino acids likely block the substrate binding site, leading to a nonfunctional enzyme.

BTZ043, likely because two or three bases of the cysteine codon inhibition in all the mutant enzymes. This residual enzymatic ac- (TGC) need to be mutated to generate a threonine (ACA, ACC, tivity explains the ability of the mutant to survive in vitro even in and ACT) or an asparagine (AAT) codon. the presence of the compound at concentrations much higher The fact that C387S and C387A mutations do not result in M. than the MIC. In the case of the noncovalent DprE1 inhibitor, tuberculosis growth defects or in reduced enzymatic capabilities C387G and C387N mutations modestly affect its activity although and are predicted to be nondeleterious mutations that do not to a considerably lesser extent than that of PBTZ169. These find- block the access of the substrate to the binding pocket are consis- ings are further supported by WT and mutant DprE1 structures tent with their occurrence in nature. Although C387G and C387N docked with either PTBZ169 or Ty38c, where the effects of C387 mutations appear to allow substrate access in structural models, it mutations on the covalent and noncovalent mechanisms of inhi- appears that the nonoptimal size of their side chains (i.e., either bition are modeled at the atomic level. too small or too large) strongly impacts the resulting shape of the With the initial BTZ-resistant mutant selection procedure, substrate binding pocket and would affect the stability and bind- mutants conferring lower-level resistance may have been missed ing of the substrate, resulting in reduced enzymatic activity. Fur- due to growth defects. An explanation for the unsuccessful selec- thermore, these mutations are also predicted to be deleterious and tion of BTZ-resistant mutants in M. smegmatis may be due to the would have a detrimental impact on the biological function of differences in the regulation of arabinogalactan or lipoarabino- DprE1. As seen experimentally, growth defects were observed on mannan biosynthesis in the two species or to the different solid medium in the presence of these mutations. This could be a genomic organization of M. smegmatis. Furthermore, it was re- consequence of the impaired catalytic capability of DprE1, which ported previously that the level of activity of the M. tuberculosis would imply that the epimerization of DPR to DPA and, by exten- embA promoter was markedly lower in M. smegmatis (33), sug- sion, the formation of the arabinogalactan layer in the mycobac- gesting that control of gene expression may be organized differ- terium cell wall occur at lower rates. ently in this nonpathogenic mycobacterium. Therefore, it appears DprE1 C387 mutations are less cytotoxic than the WT in mac- that M. smegmatis may not be the best model for studying genes rophages. Regardless of the substitution at C387, there is an im- implicated in arabinan-derived cell wall components expressed paired ability of M. tuberculosis to induce host cell death. The exact under the control of M. tuberculosis promoters. mechanism remains unclear; however, one might speculate that In conclusion, we have identified five mutations in DprE1 that these mutations result in slower replication of the mycobacterium do not affect the viability of M. tuberculosis, which give rise to a in an ex vivo context. Another possibility is that the M. tuberculosis functional enzyme and generate BTZ resistance. As these five cell wall component LAM is formed at a reduced rate as a result of BTZ-resistant mutants appear less cytotoxic in macrophages, they decreased DPA synthesis. Since LAM has been found to inhibit may also be less fit should these mutations arise in humans. From phagosome maturation within macrophages (32), this capability a clinical standpoint, an understanding of DprE1-mediated BTZ could be reduced in the presence of C387 mutations, with a con- resistance will facilitate clinical trials of the drug candidate sequent decrease in the survival of the pathogen in the host cell. PBTZ169 by screening M. tuberculosis for C387 mutations over The importance of the C387 residue in the binding of covalent the course of treatment. In the case of the C387N mutant with an DprE1 inhibitors is illustrated biochemically by the increase in intermediate level of resistance to PBTZ169, it appears that the

PBTZ169 IC50 values in the presence of C387 substitutions and the synergy between PBTZ169 and BDQ does not depend on the co- fact that even at high concentrations of PBTZ169, there is no full valent bond of C387 with PBTZ169. This is consistent with the

November 2016 Volume 60 Number 11 Antimicrobial Agents and Chemotherapy aac.asm.org 6457 167 Foo et al.

finding that noncovalent DprE1 inhibitors, the 1,4-azaindoles, benzothiazinone-mediated killing of Mycobacterium tuberculosis. Sci also display synergism with BDQ (34). The lack of cross-resistance Transl Med 4:150ra121. http://dx.doi.org/10.1126/scitranslmed.3004395. of BTZ-resistant mutants to Ty38c as reported in this work would 10. Batt SM, Jabeen T, Bhowruth V, Quill L, Lund PA, Eggeling L, Alder- wick LJ, Futterer K, Besra GS. 2012. Structural basis of inhibition of also be an important consideration when deciding which and how Mycobacterium tuberculosis DprE1 by benzothiazinone inhibitors. Proc many DprE1 inhibitors to develop clinically, as a noncovalent Natl Acad SciUSA109:11354–11359. http://dx.doi.org/10.1073/pnas DprE1 inhibitor could be used after resistance to the covalent one .1205735109. develops, or vice versa. Altogether, this work provides timely in- 11. Brecik M, Centárová I, Mukherjee R, Kolly GS, Huszár S, Bobovská A, Kilacsková E, Mokošová V, Svetlikova Z, Šarkan M, Neres J, Kordu- sight into the mechanisms of resistance of M. tuberculosis to láková J, Cole ST, Mikušová K. 29 April 2015. DprE1 is a vulnerable DprE1 inhibitors as PBTZ169 progresses into clinical trials. tuberculosis drug target due to its cell wall localization. ACS Chem Biol http://dx.doi.org/10.1021/acschembio.5b00237. ACKNOWLEDGMENTS 12. Magnet S, Hartkoorn RC, Székely R, Pató J, Triccas JA, Schneider P, Szántai-Kis C, O˝ rfi L, Chambon M, Banfi D, Bueno M, Turcatti G, Kéri We thank Neeraj Dhar for providing plasmid pND255, Vadim Makarov G, Cole ST. 2010. Leads for antitubercular compounds from kinase in- for BTZ043 and PBTZ169, Maria-Paola Costi for Ty38c, and Anthony hibitor library screens. Tuberculosis 90:354–360. http://dx.doi.org/10 Vocat for his technical assistance in REMAs. .1016/j.tube.2010.09.001. 13. Stanley SA, Grant SS, Kawate T, Iwase N, Shimizu M, Wivagg C, Silvis FUNDING INFORMATION M, Kazyanskaya E, Aquadro J, Golas A, Fitzgerald M, Dai H, Zhang L, Hung DT. 2012. Identification of novel inhibitors of M. tuberculosis This work was supported by the European Community’s Seventh Frame- growth using whole cell based high-throughput screening. ACS Chem Biol work Program FP7/2007-2013 under grant agreement 260872. Benoit 7:1377–1384. http://dx.doi.org/10.1021/cb300151m. Lechartier was a recipient of a grant from the Fondation Jacqueline Bey- 14. Wang F, Sambandan D, Halder R, Wang J, Batt SM, Weinrick B, tout. João Neres was awarded a Marie Curie fellowship. Ahmad I, Yang P, Zhang Y, Kim J, Hassani M, Huszar S, Trefzer C, Ma Z, Kaneko T, Mdluli KE, Franzblau S, Chatterjee AK, Johnsson K, Mikusova K, Besra GS, Fütterer K, Robbins SH, Barnes SW, Walker JR, REFERENCES Jacobs WR, Schultz PG. 2013. Identification of a small molecule with 1. WHO. 2015. Global tuberculosis report 2015. WHO, Geneva, Switzer- activity against drug-resistant and persistent tuberculosis. Proc Natl Acad land. Sci U S A 110:E2510–E2517. http://dx.doi.org/10.1073/pnas.1309171110. 2. Makarov V, Manina G, Mikusova K, Mollmann U, Ryabova O, Saint- 15. Shirude PS, Shandil R, Sadler C, Naik M, Hosagrahara V, Hameed S, Joanis B, Dhar N, Pasca MR, Buroni S, Lucarelli AP, Milano A, De Shinde V, Bathula C, Humnabadkar V, Kumar N, Reddy J, Panduga V, Rossi E, Belanova M, Bobovska A, Dianiskova P, Kordulakova J, Sala C, Sharma S, Ambady A, Hegde N, Whiteaker J, McLaughlin RE, Gardner Fullam E, Schneider P, McKinney JD, Brodin P, Christophe T, Waddell H, Madhavapeddi P, Ramachandran V, Kaur P, Narayan A, Guptha S, S, Butcher P, Albrethsen J, Rosenkrands I, Brosch R, Nandi V, Bharath Awasthy D, Narayan C, Mahadevaswamy J, Vishwas KG, Ahuja V, S, Gaonkar S, Shandil RK, Balasubramanian V, Balganesh T, Tyagi S, Srivastava A, Prabhakar K, Bharath S, Kale R, Ramaiah M, Choudhury Grosset J, Riccardi G, Cole ST. 2009. Benzothiazinones kill Mycobacte- NR, Sambandamurthy VK, Solapure S, Iyer PS, Narayanan S, Chatterji rium tuberculosis by blocking arabinan synthesis. Science 324:801–804. M. 2013. Azaindoles: noncovalent DprE1 inhibitors from scaffold mor- http://dx.doi.org/10.1126/science.1171583. phing efforts, kill Mycobacterium tuberculosis and are efficacious in vivo. 3. Pasca MR, Degiacomi G, de Jesus Lopes Ribeiro AL, Zara F, De Mori J Med Chem 56:9701–9708. http://dx.doi.org/10.1021/jm401382v. P, Heym B, Mirrione M, Brerra R, Pagani L, Pucillo L, Troupioti P, 16. Panda M, Ramachandran S, Ramachandran V, Shirude PS, Humna- Makarov V, Cole ST, Riccardi G. 2010. Clinical isolates of Mycobacte- badkar V, Nagalapur K, Sharma S, Kaur P, Guptha S, Narayan A, rium tuberculosis in four European hospitals are uniformly susceptible to Mahadevaswamy J, Ambady A, Hegde N, Rudrapatna SS, Hosagrahara benzothiazinones. Antimicrob Agents Chemother 54:1616–1618. http: VP, Sambandamurthy VK, Raichurkar A. 2014. Discovery of pyrazol- //dx.doi.org/10.1128/AAC.01676-09. opyridones as a novel class of non-covalent dpre1 inhibitor with potent 4. Makarov V, Lechartier B, Zhang M, Neres J, van der Sar AM, Raadsen anti-mycobacterial activity. J Med Chem 57:4761–4771. http://dx.doi.org SA, Hartkoorn RC, Ryabova OB, Vocat A, Decosterd LA, Widmer N, /10.1021/jm5002937. Buclin T, Bitter W, Andries K, Pojer F, Dyson PJ, Cole ST. 2014. 17. Naik M, Humnabadkar V, Tantry SJ, Panda M, Narayan A, Guptha S, Towards a new combination therapy for tuberculosis with next generation Panduga V, Manjrekar P, Jena LK, Koushik K, Shanbhag G, Jatheen- benzothiazinones. EMBO Mol Med 6:372–383. http://dx.doi.org/10.1002 dranath S, Manjunatha MR, Gorai G, Bathula C, Rudrapatna S, Achar /emmm.201303575. V, Sharma S, Ambady A, Hegde N, Mahadevaswamy J, Kaur P, Sam- 5. Working Group on New TB Drugs. 2015. Drug pipeline. Stop TB Part- bandamurthy VK, Awasthy D, Narayan C, Ravishankar S, Madhavape- nership, Geneva, Switzerland. http://www.newtbdrugs.org/pipeline.php. ddi P, Reddy J, Prabhakar K, Saralaya R, Chatterji M, Whiteaker J, Accessed 22 November 2015. McLaughlin B, Chiarelli LR, Riccardi G, Pasca MR, Binda C, Neres J, 6. Mikusova K, Huang H, Yagi T, Holsters M, Vereecke D, D’Haeze W, Dhar N, Signorino-Gelo F, McKinney JD, Ramachandran V, Shandil R, Scherman MS, Brennan PJ, McNeil MR, Crick DC. 2005. Decaprenyl- Tommasi R, Iyer PS, Narayanan S, Hosagrahara V, Kavanagh S, Dinesh phosphoryl arabinofuranose, the donor of the D-arabinofuranosyl resi- N, Ghorpade SR. 2014. 4-Aminoquinolone piperidine amides: noncova- dues of mycobacterial arabinan, is formed via a two-step epimerization of lent inhibitors of DprE1 with long residence time and potent antimyco- decaprenylphosphoryl ribose. J Bacteriol 187:8020–8025. http://dx.doi bacterial activity. J Med Chem 57:5419–5434. http://dx.doi.org/10.1021 .org/10.1128/JB.187.23.8020-8025.2005. /jm5005978. 7. Trefzer C, Rengifo-Gonzalez M, Hinner MJ, Schneider P, Makarov V, 18. Neres J, Hartkoorn RC, Chiarelli LR, Gadupudi R, Pasca MR, Mori G, Cole ST, Johnsson K. 2010. Benzothiazinones: prodrugs that covalently Venturelli A, Savina S, Makarov V, Kolly GS, Molteni E, Binda C, Dhar modify the decaprenylphosphoryl-␤-D-ribose 2=-epimerase DprE1 of N, Ferrari S, Brodin P, Delorme V, Landry V, de Jesus Lopes Ribeiro Mycobacterium tuberculosis. J Am Chem Soc 132:13663–13665. http://dx AL, Farina D, Saxena P, Pojer F, Carta A, Luciani R, Porta A, Zanoni .doi.org/10.1021/ja106357w. G, De Rossi E, Costi MP, Riccardi G, Cole ST. 2015. 2-Carboxyqui- 8. Trefzer C, Škovierová H, Buroni S, Bobovská A, Nenci S, Molteni E, noxalines kill Mycobacterium tuberculosis through noncovalent inhibi- Pojer F, Pasca MR, Makarov V, Cole ST, Riccardi G, Mikušová K, tion of DprE1. ACS Chem Biol 10:705–714. http://dx.doi.org/10.1021 Johnsson K. 2012. Benzothiazinones are suicide inhibitors of mycobac- /cb5007163. terial decaprenylphosphoryl-␤-D-ribofuranose 2=-oxidase DprE1. J Am 19. Palomino J-C, Martin A, Camacho M, Guerra H, Swings J, Portaels F. Chem Soc 134:912–915. http://dx.doi.org/10.1021/ja211042r. 2002. Resazurin microtiter assay plate: simple and inexpensive method for 9. Neres J, Pojer F, Molteni E, Chiarelli LR, Dhar N, Boy-Rottger S, detection of drug resistance in Mycobacterium tuberculosis. Antimicrob Buroni S, Fullam E, Degiacomi G, Lucarelli AP, Read RJ, Zanoni G, Agents Chemother 46:2720–2722. http://dx.doi.org/10.1128/AAC.46.8 Edmondson DE, De Rossi E, Pasca MR, McKinney JD, Dyson PJ, .2720-2722.2002. Riccardi G, Mattevi A, Cole ST, Binda C. 2012. Structural basis for 20. van Kessel JC, Marinelli LJ, Hatfull GF. 2008. Recombineering myco-

6458 aac.asm.org Antimicrobial Agents and Chemotherapy November 2016 Volume 60 Number 11 168 Mycobacterial DprE1 and BTZ Resistance

bacteria and their phages. Nat Rev Microbiol 6:851–857. http://dx.doi.org 29. Lechartier B, Hartkoorn RC, Cole ST. 2012. In vitro combination studies /10.1038/nrmicro2014. of benzothiazinone lead compound BTZ043 against Mycobacterium tu- 21. Rybniker J, Vocat A, Sala C, Busso P, Pojer F, Benjak A, Cole ST. 2015. berculosis. Antimicrob Agents Chemother 56:5790–5793. http://dx.doi Lansoprazole is an antituberculous prodrug targeting cytochrome bc1. .org/10.1128/AAC.01476-12. Nat Commun 6:7659. http://dx.doi.org/10.1038/ncomms8659. 30. Riccardi G, Pasca MR, Chiarelli LR, Manina G, Mattevi A, Binda C. 22. Van Durme J, Delgado J, Stricher F, Serrano L, Schymkowitz J, Rous- 2013. The DprE1 enzyme, one of the most vulnerable targets of Mycobac- seau F. 2011. A graphical interface for the FoldX forceﬁeld. Bioinformat- terium tuberculosis. Appl Microbiol Biotechnol 97:8841–8848. http://dx ics 27:1711–1712. http://dx.doi.org/10.1093/bioinformatics/btr254. .doi.org/10.1007/s00253-013-5218-x. 23. Krieger E, Vriend G. 2014. YASARA View—molecular graphics for all 31. Mikusova K, Makarov V, Neres J. 2013. DprE1—from the discovery to devices—from smartphones to workstations. Bioinformatics 30:2981– the promising tuberculosis drug target. Curr Pharm Des 20:4379–4403. 2982. http://dx.doi.org/10.1093/bioinformatics/btu426. http://dx.doi.org/10.2174/138161282027140630122724. 24. Schrödinger LLC. 2015. The PyMOL molecular graphics system, version 32. Vergne I, Fratti RA, Hill PJ, Chua J, Belisle J, Deretic V. 2004. Myco- 1.8. Schrödinger, LLC, Cambridge, MA. bacterium tuberculosis phagosome maturation arrest: mycobacterial 25. Choi Y, Sims GE, Murphy S, Miller JR, Chan AP. 2012. Predicting the phosphatidylinositol analog phosphatidylinositol mannoside stimulates functional effect of amino acid substitutions and indels. PLoS One early endosomal fusion. Mol Biol Cell 15:751–760. http://dx.doi.org/10 7:e46688. http://dx.doi.org/10.1371/journal.pone.0046688. .1091/mbc.E03-05-0307. 26. Choi Y, Chan AP. 2015. PROVEAN Web server: a tool to predict the 33. Amin AG, Goude R, Shi L, Zhang J, Chatterjee D, Parish T. 2008. functional effect of amino acid substitutions and indels. Bioinformatics EmbA is an essential arabinosyltransferase in Mycobacterium tubercu- 31:2745–2747. http://dx.doi.org/10.1093/bioinformatics/btv195. losis. Microbiology 154:240–248. http://dx.doi.org/10.1099/mic.0 27. Kolly GS, Mukherjee R, Kilacsková E, Abriata LA, Raccaud M, Blaško .2007/012153-0. J, Sala C, Dal Peraro M, Mikušová K, Cole ST. 2015. GtrA protein 34. Chatterji M, Shandil R, Manjunatha MR, Solapure S, Ramachandran V, Rv3789 is required for arabinosylation of arabinogalactan in Mycobacte- Kumar N, Saralaya R, Panduga V, Reddy J, Prabhakar KR, Sharma S, rium tuberculosis. J Bacteriol 197:3686–3697. http://dx.doi.org/10.1128 Sadler C, Cooper CB, Mdluli K, Iyer PS, Narayanan S, Shirude PS. /JB.00628-15. 2014. 1,4-Azaindole, a potential drug candidate for treatment of tubercu- 28. Hsu T, Hingley-Wilson SM, Chen B, Chen M, Dai AZ, Morin PM, losis. Antimicrob Agents Chemother 58:5325–5331. http://dx.doi.org/10 Marks CB, Padiyar J, Goulding C, Gingery M. 2003. The primary .1128/AAC.03233-14. mechanism of attenuation of bacillus Calmette-Guerin is a loss of se- 35. Lambert RJW, Pearson J. 2000. Susceptibility testing: accurate and re- creted lytic function required for invasion of lung interstitial tissue. producible minimum inhibitory concentration (MIC) and noninhibitory Proc Natl Acad SciUSA100:12420–12425. http://dx.doi.org/10.1073 concentration (NIC) values. Journal of Applied Microbiology 88:784– /pnas.1635213100. 790. http://dx.doi.org/10.1046/j.1365-2672.2000.01017.x.

November 2016 Volume 60 Number 11 Antimicrobial Agents and Chemotherapy aac.asm.org 6459 169 6833/(0(17$/0$7(5,$/

SUPPLEMENTARYMETHODS

Determination of PBTZ169 and BDQ interaction in H37Rv and C387N mutant. Drug interactionsbetweenPBTZ169andBDQweredeterminedusingthecheckerboardassay[1,2].

M.tuberculosisstrainsweregrownin7H9mediumtolog-phase(OD600 0.4-0.8)anddiluted

3 toanOD600 of0.0001. 75μlofbacterialsuspension(2.25x10 cells)wasaddedperwellof a96-wellplate. Thefirstcompoundwasaddedtothefirstcolumnandtwo-foldserial dilutionsweremadecolumn-wise.Thesecondcompoundwaspreparedin7H9ata concentration8xMICandserialdilutionsweremadeto0.125xMIC.25μlofdiluted compoundateachconcentrationwasaddedtoarowofthe96-wellplate.Variouscombinations ofPBTZ169andBDQwerethusobtainedinafinalvolumeof100μlperwell. Plateswere incubatedfor6daysat37oC,10μlof0.025%w/vresazurinwasaddedtoeachwell. The fluorescenceintensitywasreadafter24hincubationusinganInfiniteF200Tecanplate reader.

ForrowswhereanMICvaluecouldbedetermined,thefractionalinhibitoryconcentration index(ΣFICI)wascalculatedusingtheequationΣFICindex=FICPBTZ +FICBDQ =(MICof

PBTZ,testedincombination)/(MICofPBTZ,alone)+(MICofBDQ,testedincombina-tion)/

(MICofBDQ,alone).ΣFIC≤0.5isconsideredassynergism,0.5<ΣFIC≤4additivity,and

ΣFIC>4antagonism.

170 SUPPLEMENTARY FIGURES

FIG S1 Growth comparison of H37Rv and mutant strains in liquid medium with or without BTZ043. M. tuberculosis H37Rv (wt) was grown with (+BTZ) or without 400 ng/ml BTZ043. Wild-type growth was compared to the merodiploid strain harbouring the cysteine codon (C387), which behaves like the wild-type strain, and the glycine or serine mutants (C387G and C387S), which are both fully resistant to BTZ. The C387A, C387N and C387T mutants behaved like the other mutants (not shown in the graph).

FIG S2 Growth comparison between C387T and C387G mutant strains on solid media with or without BTZ043. The mutants DprE1C387G and DprE1C387T were plated with or without 40 or 400 ng/ml BTZ on the same day. The picture was taken after 21 days of incubation at 37°C. The C387G mutant CFUs on BTZ-containing plates were hardly detectable after 3 weeks of incubation but were countable after 6 weeks.

171 FIG S3 Interactions of PBTZ169 and BDQ in H37Rv and C387N mutant. The sum of fractional inhibitory concentration (FIC) of PBTZ169 and of BDQ were calculated with the equation ΣFIC index = FICPBTZ169 + FICBDQ = (MIC of PBTZ169, tested in combination)/(MIC of PBTZ169, alone) + (MIC of BDQ, tested in combination)/(MIC of BDQ, alone). ΣFIC indices were calculated for drug concentrations where an MIC could be determined using a checkerboard assay with REMA as a viability marker. Black circles represent H37Rv, grey squares represent

C387N mutant strain. The dotted line indicates the threshold for synergism. Data from two independent experiments are presented as mean ± SD.

FIG S4 Models of binding of PBTZ169 and Ty38c in the binding pocket of the ﬁve DprE1 mutants. The 387 position of DprE1 is represented in red, FAD in blue. PBTZ169 and Ty38c are shown docked into the binding pocket of DprE1 WT and mutant crystal structures. Both compounds are able to bind to the 5 mutants.

172 TABLE S1 Primers used in this study

Primer name Sequence Mut90_fwd CCATCCCGGGCTGGAACATCNNNGTCGACTTCCCCATCAAGGACG§ Mut90_rev CGTCCTTGATGGGGAAGTCGACNNNGATGTTCCAGCCCGGGATGG 90C387G_fwd CCATCCCGGGCTGGAACATCGGCGTCGACTTCCCCATCAAGGACG 90C387G_rev CGTCCTTGATGGGGAAGTCGACGCCGATGTTCCAGCCCGGGATGG Rv3790_fwd AATACTCCATGGCCATCCTGACGGATGGCCTTGCAGCCCACTAGGTCAGGC* Rv3790_rev AATGCAAGTACTCTACAGCAGCTCCAAGCGTC* JN C387S _fwd GGCTGGAACATCAGCGTCGACTTCCCC JN C387S _rev CCGACCTTGTAGTCGCAGCTGAAGGGG JN C387G _fwd GGCTGGAACATCGGCGTCGACTTCCCC JN C387G _rev CCGACCTTGTAGCCGCAGCTGAAGGGG Rv3790c387G_For gccgctcagcttccccatcccgggctggaacatcggcgtcgacttccccatcaaggacgggctggggaag Rv3790c387G_Rev cttccccagcccgtccttgatggggaagtcgacgccgatgttccagcccgggatggggaagctgagcggc Rv3790c387S_For gccgctcagcttccccatcccgggctggaacatctccgtcgacttccccatcaaggacgggctggggaag Rv3790c387S_Rev cttccccagcccgtccttgatggggaagtcgacggagatgttccagcccgggatggggaagctgagcggc Rv3790c387A_For gccgctcagcttccccatcccgggctggaacatcgccgtcgacttccccatcaaggacgggctggggaag Rv3790c387A_Rev cttccccagcccgtccttgatggggaagtcgacggcgatgttccagcccgggatggggaagctgagcggc Rv3790c387T_For gccgctcagcttccccatcccgggctggaacatcaccgtcgacttccccatcaaggacgggctggggaag Rv3790c387T_Rev cttccccagcccgtccttgatggggaagtcgacggtgatgttccagcccgggatggggaagctgagcggc Rv3790c387N_For gccgctcagcttccccatcccgggctggaacatcaatgtcgacttccccatcaaggacgggctggggaag Rv3790c387N_Rev cttccccagcccgtccttgatggggaagtcgacattgatgttccagcccgggatggggaagctgagcggc Rv3790C387N_F CGGGCTGGAACATCAACGTCGACTTCCCCATC Rv3790C387N_R GATGGGGAAGTCGACGTTGATGTTCCAGCCCG Rv3790C387A_F CGGGCTGGAACATCGCCGTCGACTTCCCCATC Rv3790C387A_R GATGGGGAAGTCGACGGCGATGTTCCAGCCCG Rv3790C387T_F CGGGCTGGAACATCACCGTCGACTTCCCCATC Rv3790C387T_R GATGGGGAAGTCGACGGTGATGTTCCAGCCCG * Restriction sites are indicated in italics in the primer sequence § NNN: random bases used for C387 mutagenesis

TABLE S2 Plasmids used in this study

Name Description and features Resistance References marker pND255 L5-integrative vector HygR N. Dhar, unpublished pBL2 wild-type promoter –rv3789-dprE1 cloned into KanR this study pCR®– Blunt II – TOPO® Vector (Invitrogen) pET28a-dprE1 Expression vector for DprE1 protein production KanR, CamR [3] pND255-derivative plasmid pBLX1-6 Pools of vectors carrying randomly mutated HygR this study dprE1 and expressed under the natural promoter located upstream of rv3789 pBLG Mutated dprE1(C387G) and expressed under the HygR this study natural promoter located upstream of rv3789

173 TABLE S3 Strains used in this study

Strain Plasmid Genotype Reference H37Rv - Wild-type [4] H37Rv::dprE1 (C387X) pBLX1-6* wt promoter-rv3789- dprE1(C387X) at the this study L5-attB site H37Rv::dprE1 (C387G) pBLG wt promoter-rv3789- dprE1(C387G) at the this study L5-attB site H37Rv/pJV53 pJV53 Expression of Che9c gene products 60 and [5] 61 to facilitate double-stranded DNA re- combination in mycobacteria H37RvΔRD1 - Deletion of RD1 region of H37Rv [6] dprE1 (C387S) - H37Rv strain carrying C387S mutation in this study dprE1 dprE1 (C387A) - H37Rv strain carrying C387A mutation in this study dprE1 dprE1 (C387T) - H37Rv strain carrying C387T mutation in this study dprE1 dprE1 (C387G) - H37Rv strain carrying C387G mutation in this study dprE1 dprE1 (C387N) - H37Rv strain carrying C387N mutation in this study dprE1

174 REFERENCES

1. Reddy, V. M., Einck, L., Andries, K., and Nacy, C. A. 2010. In Vitro Interactions

between New Antitubercular Drug Candidates SQ109 and TMC207. Antimicrobial Agents

and Chemotherapy, 54(7):2840–2846. doi:10.1128/AAC.01601-09.

2. Lechartier, B., Hartkoorn, R. C., and Cole, S. T. 2012. In Vitro Combination Studies of

Benzothiazinone Lead Compound BTZ043 against Mycobacterium tuberculosis. Antimicro-

bial Agents and Chemotherapy, 56(11):5790–5793. doi:10.1128/AAC.01476-12.

3. Makarov, V., Lechartier, B., Zhang, M., Neres, J., van der Sar, A. M., Raadsen, S. A.,

Hartkoorn, R. C., Ryabova, O. B., Vocat, A., Decosterd, L. A., Widmer, N., Buclin,

T., Bitter, W., Andries, K., Pojer, F., Dyson, P. J., and Cole, S. T. 2014. Towards a

new combination therapy for tuberculosis with next generation benzothiazinones. EMBO

Molecular Medicine, 6(3):372–383. doi:10.1002/emmm.201303575.

4. Cole, S. T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., Gordon, S. V.,

Eiglmeier, K., Gas, S., Barry, C. E., Tekaia, F., Badcock, K., Basham, D., Brown, D.,

Chillingworth, T., Connor, R., Davies, R., Devlin, K., Feltwell, T., Gentles, S., Hamlin,

N., Holroyd, S., Hornsby, T., Jagels, K., Krogh, A., McLean, J., Moule, S., Murphy,

L., Oliver, K., Osborne, J., Quail, M. A., Rajandream, M.-A., Rogers, J., Rutter, S.,

Seeger, K., Skelton, J., Squares, R., Squares, S., Sulston, J. E., Taylor, K., Whitehead,

S., and Barrell, B. G. 1998. Deciphering the biology of Mycobacterium tuberculosis from

the complete genome sequence. Nature, 393(6685):537–544. doi:10.1038/31159.

175 5. van Kessel, J. C., Marinelli, L. J., and Hatfull, G. F. 2008. Recombineering mycobacteria

and their phages. Nature Reviews Microbiology, 6(11):851–857.

6. Hsu, T., Hingley-Wilson, S. M., Chen, B., Chen, M., Dai, A. Z., Morin, P. M., Marks,

C. B., Padiyar, J., Goulding, C., and Gingery, M. 2003. The primary mechanism of atten-

uation of bacillus Calmette–Guerin is a loss of secreted lytic function required for invasion

of lung interstitial tissue. Proceedings of the National Academy of Sciences, 100(21):12420–

12425.

176 Chapter 5: Conclusions and Perspectives

177

Figure 1: Overview of the main contributions of this thesis. Discovery of (A) AX compounds (Chapter 2), (B) PB compounds (Chapter 3), and (C) clinical development support of BTZs (Chapter 4, Annex Chapter A1, A2, and A3).

179 To fulfil the aim of this thesis, which has been to enhance the global TB drug pipeline in order to contribute to new drugs active against M. tuberculosis, the discovery of two new, highly promising families of anti-tuberculars, the arylvinylpiperazine amides AX in Chapter 2 and the lapachol-derived PB compounds in Chapter 3, as well as the elucidation of the underlying mechanisms of resistance to BTZs in Chapter 4, have been presented in the earlier chapters of this thesis.

Chapter 2 details the characterisation and extensive profiling of the AX compounds from a whole cell-based approach through activities associated with hit-to-lead generation and lead optimisation (Fig. 1a). This family remains interesting due to their potent activity against M. tuberculosis and their ease of chemical synthesis. While AX-35 is the lead compound in vitro, structural modifications to this parent molecule improved activity of two analogs tested in vivo.

Future work should therefore be focused on advancing this family, or at least these three more potent analogs, further in the lead optimisation phase. This would imply improving liabilities of these compounds, one of which identified in this study is their metabolic stability, which is likely to influence their in vivo efficacy and should be confirmed by future pharmacokinetics studies in mice. In this respect, biochemical information obtained in this work with regards to important residues for target binding could guide future structure-based SAR studies.

The target of AX compounds was identified in this thesis as the QcrB subunit of the cytochrome bc1-aa3 complex. Since AX compounds have a different mode of interaction in the same binding site in QcrB as Q203, these compounds could be further developed as backup QcrB inhibitors, in particular in the event of emergence of resistance to Q203. That QcrB is the target of several structurally diverse chemical entities may be attributed to its localisation in the mycobacterial membrane, similar to other promiscuous targets such as DprE1 and MmpL3,

180 where its active site is exposed to the periplasm (1). As it often occurs with membrane proteins, the expression and purification of functional enzymes is challenging and to date, a purified, functional QcrB has yet to be reported. Without the purified enzyme, biochemical validation of the target and target-based SAR studies are both limited. In the case of AX compounds, this study has validated the target as QcrB through alternative, multi-faceted approaches of genetics, transcriptomics, and bioenergetic flux assays. Gene expression profiling studies have also reaffirmed the compensatory respiratory role of the alternate terminal oxidase, cytochrome bd.

With inhibition of QcrB and in the absence of cytochrome bd oxidase, AX compounds become bactericidal. Therefore, future work on AX compounds should also encompass combination studies with potential small-molecule inhibitors of cytochrome bd oxidase, as demonstrated with aurachin D (2).

Ironically, this respiratory flexibility between the two terminal oxidases can also be used to the disadvantage of the bacillus, as demonstrated by the potent killing of M. tuberculosis with a combination of BDQ, Q203, and CFM, whereby activation of cytochrome bd oxidase upon inhibition of cytochrome bc1-aa3 potentiates CFM activity (3). Therefore, dysregulation of the terminal oxidases represents a novel strategy against the bacillus, and could be incorporated into a new anti-TB regimen mainly targeting oxidative phosphorylation. Such a regimen would extensively perturb respiratory flux and ATP production, and theoretically would be able to target the spectrum of metabolic states of M. tuberculosis physiologically and shorten treatment duration, particularly as ATP homeostasis is critical for sustaining all states of the bacillus (4).

For such a regimen to be effectively implemented however, these drugs must be highly specific for the mycobacterial respiratory components without affecting those of the human mitochondrial counterparts. However, it is highly likely that a mutation in any of the targets would threaten the efficacy of the other compounds within the regimen, and therefore such a

181 combination may be more suited for MDR- and XDR-TB cases rather than as a front-line treatment.

In Chapter 3, the characterisation and profiling of lapachol-derived PB compounds through a whole cell-based approach with mainly hit-to-lead generation activities are elaborated (Fig. 1b).

PB analogs are interesting as anti-tuberculars in that they derive from a naturally-occurring, readily-available source, are easily synthesised, and have substantially improved activity against M. tuberculosis compared to the parent scaffold, lapachol. Furthermore, these molecules demonstrate activity against non-replicating M. tuberculosis. Future work should focus on advancing this family further through lead optimisation activities, and prioritising PB-089, PB-

117, and PB-118 analogs in particular, since PB-116 does not demonstrate intracellular activity in infected macrophages. A liability of this family identified in this study is their metabolic instability, which can be addressed through more medicinal chemistry studies. Additional key activities should include pharmacokinetics studies in vivo, as well as determining the in vivo efficacy of these compounds in both replicating and non-replicating models of TB.

This study establishes the reliance on the F420 cofactor for activity of PB analogs against M. tuberculosis. While the F420-dependent mechanism of activity is similar to Delamanid and PA-

824, what is different in the case of the PB compounds is that Ddn is not required for activity, and therefore the prospect of a novel, mainly MTBC-specific mechanism of action for this family is exciting. The discovery program of PB compounds should continue independently, and these molecules could additionally serve as alternatives for Delamanid and PA-824 should phenotypic resistance, mediated by mutations in Ddn, arise to both these agents. To obtain more insight into the mechanism of action of PB compounds, the next steps would be to identify the

F420-dependent activator and the highly active species to which PB analogs are activated to. The

182 possibility of different mechanisms being attributed to the activity of PB compounds against replicating and non-replicating M. tuberculosis should also be explored, as with Delamanid and

PA-824.

Since there is an unusually high dependency of M. tuberculosis on the F420 cofactor (5), which may be required for its survival under anaerobic conditions (6), the F420 biosynthesis pathway could be a novel, attractive target against non-replicating M. tuberculosis for shortening TB treatment durations. The relevance of this pathway as a drug target remains to be validated in vivo, and should it prove essential, a target-based, whole-cell screening approach could be undertaken to identify hits targeting F420 biosynthesis by measuring the resistance of M. tuberculosis to PB, Delamanid, or PA-824.

Lastly, this thesis provides support to the ongoing pre-clinical and clinical development of

BTZs. The underlying resistance mechanisms of M. tuberculosis to these highly potent compounds is elucidated in Chapter 4 (Fig. 1c). This study identifies 5 out of 19 possible mutations occurring at the cysteine 387 residue of DprE1 mediating resistance to BTZs, two of which are reported for the first time here. Apart from the identification of BTZ-resistance markers, profiling of the levels of resistance for these mutations is also important clinically as it can guide decisions in the inclusion and dosage of BTZs as part of a regimen. This study also reveals that all five mutations impose fitness costs on M. tuberculosis intracellularly, indicating that these mutations, and consequently, BTZ-resistance, may have a lower tendency to emerge clinically. This hypothesis needs to be verified by monitoring the frequency of phenotypic resistance to PBTZ169 in patients and screening for the five C387 mutations of DprE1 over the course of clinical trials. The lack of cross-resistance between covalent and non-covalent DprE1 inhibitors biochemically is further highlighted in this study. Differences between these two

183 classes in interactions within the binding pocket of DprE1 can further inspire and rationally guide structure-based design to generate potential backup molecules and series of BTZs to overcome the emergence of BTZ-resistance, as demonstrated by the generation of sulfonamide

PBTZs (Annex A1, A2). In addition, since combination therapy is required for TB treatment to reduce the emergence of antibiotic resistance and to target multiple sub-populations of bacilli, new TB drug candidates need to be evaluated as part of regimens. As demonstrated in Annex

A3, MCZ does not display antagonistic interactions with conventional and new TB drugs, demonstrating its compatibility as a component of a novel anti-TB regimen.

In a broader context, several observations and considerations related to TB drug discovery and development can be made from the course of this thesis and other encountered works. Firstly, it is important to extensively characterise new molecules for their anti-tubercular activity across different systems and models, since the susceptibility of the bacillus to compounds and the physiological relevance of the targets are highly influenced by the metabolic state of M. tuberculosis. Secondly, natural product entities and human-targeted drugs represent promising alternatives to current small, synthetic molecule libraries for expanding and diversifying chemical space of anti-tuberculars with novel mechanisms of action (7). Since large, complex natural product entities often pose challenges for downstream SAR studies and optimisation, those with simpler scaffolds such as lapachol could be prioritised for TB drug discovery. With regards to re-purposing human-targeted drugs for TB, a case in point is lansoprazole (LPZ), a proton pump inhibitor requiring activation by host cells to LPZ sulfide for potent activity against M. tuberculosis (8). Thirdly, while the focus of the compounds described in this thesis has been mainly on their activity against M. tuberculosis, the therapeutic potential of these molecules (and others being developed for TB) may be expanded to other mycobacterial diseases, such as Buruli ulcer and leprosy, leading to alternative avenues for their development.

184 Another important consideration is that since anti-TB drugs are preferably orally dosed, interactions between the gut microbiome should be taken into account for therapy and drug development. As gut bacteria are capable of modulating drug efficacy and toxicity (9), characterising such drug-microbiome interactions would aid in maximising efficacy of a drug.

A further point of consideration is the risk of acquired antibiotic resistance through the consumption of human-targeted, non-antimicrobial drugs (7), and thus monitoring such events would be important for guiding implementation of TB regimens, such as in the context of the co-disease burden of TB and diabetes. Lastly, having several different anti-TB drugs with the same mechanism of action would be advantageous in addressing possibilities of undesired drug interactions, thus ensuring the compatibility of therapies within and without, such as in cases of TB-HIV co-infection.

Naturally, much work still remains to fully realise the antibiotic potential of the compounds presented in this thesis. TB drug discovery and development typically spans over 10 to 20 years, inadvertently leading to high costs associated with this long and risky process. An estimated

US$2 billion a year is required to fund TB research and development for new drugs, vaccines, and improved diagnostics, however this amount falls woefully short of the US$0.7 billion available per year (10). This lack of financial support is an important factor impeding the progress towards eradicating TB.

As a disease associated with low socio-economic status, tackling TB in today’s context necessitates the emphasis on the affordability, accessibility, and ease of use for the effective implementation of available and novel tools. Nevertheless, TB remains a disease which does not recognise political nor geographical boundaries. Overcoming this global pandemic therefore requires a combined, concerted, international effort from all levels, from governments

185 introducing effective public health policies and robust healthcare infrastructure, public and private organisations actively involved in providing financing support, researchers contributing to the improvement and generation of novel tools, to medical and healthcare workers on the front-line with TB patients. Altogether, motivated by common goals and guided by bold visions, such as, to quote Bill Gates, “Why can’t we treat TB in two weeks?”, it is through these borderless co-operations that we can envision and work towards a TB-free world.

186 References

1. Cole ST. 2016. Inhibiting Mycobacterium tuberculosis within and without. Philos Trans R

Soc B Biol Sci 371.

2. Lu P, Asseri AH, Kremer M, Maaskant J, Ummels R, Lill H, Bald D. 2018. The anti-

mycobacterial activity of the cytochrome bcc inhibitor Q203 can be enhanced by small-

molecule inhibition of cytochrome bd. Sci Rep 8:2625.

3. Lamprecht DA, Finin PM, Rahman MA, Cumming BM, Russell SL, Jonnala SR, Adamson

JH, Steyn AJC. 2016. Turning the respiratory flexibility of Mycobacterium tuberculosis

against itself. Nat Commun 7:12393.

4. Rao SPS, Alonso S, Rand L, Dick T, Pethe K. 2008. The protonmotive force is required for

maintaining ATP homeostasis and viability of hypoxic, nonreplicating Mycobacterium

tuberculosis. Proc Natl Acad Sci 105:11945–11950.

5. Selengut JD, Haft DH. 2010. Unexpected Abundance of Coenzyme F420-Dependent

Enzymes in Mycobacterium tuberculosis and Other Actinobacteria. J Bacteriol 192:5788–

5798.

6. Boshoff HIM, Barry 3rd CE. 2005. Tuberculosis — metabolism and respiration in the

absence of growth. Nat Rev Microbiol 3:70–80.

7. Maier L, Pruteanu M, Kuhn M, Zeller G, Telzerow A, Anderson EE, Brochado AR,

Fernandez KC, Dose H, Mori H, Patil KR, Bork P, Typas A. 2018. Extensive impact of

non-antibiotic drugs on human gut bacteria. Nature 555:623–628.

8. Rybniker J, Vocat A, Sala C, Busso P, Pojer F, Benjak A, Cole ST. 2015. Lansoprazole is

an antituberculous prodrug targeting cytochrome bc1. Nat Commun 6:7659.

187 9. Spanogiannopoulos P, Bess EN, Carmody RN, Turnbaugh PJ. 2016. The microbial

pharmacists within us: a metagenomic view of xenobiotic metabolism. Nat Rev Microbiol

14:273–287.

10. World Health Organization. 2017. Global Tuberculosis Report 2017. S.l.

188 Appendix Chapter A1: Structural studies of Mycobacterium tuberculosis DprE1 interacting with its inhibitors

Jérémie Piton, Caroline S.Y. Foo and Stewart T. Cole

Global Health Institute, École Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland

Drug Discovery Today, 2017, doi.org/10.1016/j.drudis.2016.09.014

Contributions: manuscript preparation

189

REVIEWS Drug Discovery Today Volume 22, Number 3 March 2017

Reviews

Structural studies of Mycobacterium

OTSCREEN POST

tuberculosis DprE1 interacting with its

inhibitors

Je´re´mie Piton, Caroline S.-Y. Foo and Stewart T. Cole

Global Health Institute, Ecole Polytechnique Fe´de´rale de Lausanne, Station 19, 1015 Lausanne, Switzerland

The ﬂavoenzyme DprE1 catalyses a crucial step in arabinan production for cell wall biosynthesis in

Mycobacterium tuberculosis and is a highly vulnerable drug target. It was ﬁrst discovered using

benzothiazinones (BTZ): exquisitely potent bactericidal agents that are being developed as drugs to treat

tuberculosis. Subsequently, many compounds with diverse scaffolds were found to act as either covalent

or noncovalent DprE1 inhibitors. Covalent inhibitors, like the BTZ, are all nitroaromatic compounds

that serve as suicide substrates after DprE1-mediated nitroreduction. Here, we describe how high-

resolution structures of DprE1, alone and in complex with various ligands, explain enzyme activity and

inhibition.

Introduction catalyses the ﬁrst step of epimerisation, namely oxidation of DPR

Mycobacteria, including Mycobacterium tuberculosis, possess a re- to an intermediary decaprenyl-phospho-2 -keto-D-arabinose

markably elaborate cell wall structure. In addition to the peptido- (DPX) with reduction of the FAD cofactor to FADH2. DPX is then

glycan layer, the mycobacterial cell wall is composed of two other reduced to DPA by the NADH-dependent decaprenylphosphoryl-

major components, namely arabinogalactan and mycolic acids, D-2-keto erythro pentose reductase (DprE2) [10]. DprE1 and DprE2

which are noncovalently linked to an outer capsule of proteins and are both essential for cell growth and the survival of M. tuberculosis

polysaccharides [1]. Some of the most effective antibiotics against [11,12]. Recently, it was found that the epimerisation process

tuberculosis (TB) to date are those targeting enzymes involved in occurs in the periplasmic space, thereby explaining the vulnera-

synthesising cell wall components. Frontline TB drugs isoniazid bility of DprE1 as a target [13].

(INH) and ethambutol (EMB) inhibit key enzymes involved in the

biosynthesis of mycolic acids and arabinan respectively [2–6]. In Inhibitors of DprE1

the past decade, whole-cell screens have enabled the identiﬁcation BTZ043, the lead compound of the benzothiazinone (BTZ) series,

of several new classes of compounds that target proteins such as was the ﬁrst DprE1 inhibitor described and is exceptionally potent

MmpL3 or DprE1, which are required for the biosynthesis of cell with an in vitro or ex vivo minimal inhibitory concentration (MIC)

well components [7,8]. These enzymes have been described as in the nanomolar range [14]. The mode of action of BTZ043

promising targets for new clinical candidates to ﬁght TB. explains its extraordinary potency because it behaves as a suicide

DprE1, encoded by the gene dprE1 (rv3790), is an essential substrate for the reduced form of DprE1. BTZ043 and other BTZ

ﬂavoenzyme of M. tuberculosis involved in cell wall synthesis derivatives undergo nitroreduction to form a nitroso derivative,

and serves as a decaprenylphosphoryl-beta-D-ribose 2 -oxidase which speciﬁcally forms a covalent adduct with cysteine 387

[9]. DprE1 acts in concert with DprE2 to catalyse the two-step (C387) in the active site of DprE1, irreversibly inhibiting the

epimerisation of decaprenyl-phospho-ribose (DPR) to decaprenyl- enzyme [15–17]. C387 is highly conserved in orthologous DprE1

phospho-arabinose (DPA), which is the sole precursor for the enzymes in actinobacteria, except in Mycobacterium avium and

synthesis of arabinogalactan and lipoarabinomannan [10]. DprE1 Mycobacterium aurum where cysteine is replaced by alanine and

serine respectively. These mutations confer natural BTZ resistance

[14].

Corresponding author: Cole, S.T. (stewart.cole@epﬂ.ch)

526 www.drugdiscoverytoday.com 191 http://dx.doi.org/10.1016/j.drudis.2016.09.014

Drug Discovery Today Volume 22, Number 3 March 2017 REVIEWS

TABLE 1

Known compounds targeting DprE1 and their inhibition properties

Covalent inhibitors Noncovalent inhibitors

Molecule Structure MIC (mM) Refs Molecule Structure MIC (mM) Refs

BTZ043 0.002 [14] TCAI 0.51 [27]

PBTZ169 0.0006 [18] TBA-7371 (1–4 azaindoles) 0.39 [31,32] POST SCREEN

DNB1 0.69 [19] Ty38c (2-carboxyquinoxalines) 3.1 [34]

Reviews

VI-9376 3.1 [20] 4-AQs (cmp-3) 0.8 [38]

377790 [21] PyrBTZ02 0.34 [25]

(8-pyrrole-BTZ)

BTO 0.07 [22] 1,2,4-Triazole containing 30 [35]

1,4-BTZ derivatives (cmp-6c)

cBT 4.2 [23] 1,3-BTZ azide 0.47 [26]

Benzothiazolylpyrimidine- 0.08 [28]

5-carboxamides (cmp-7a)

Pyrazolopyridones (cmp-19) 0.1 [33]

a 90

The IC in M. tuberculosis for this compound is 0.48 mM.

To date, 15 new classes of DprE1 inhibitors with antimycobac- and nitroimidazoles (with the lead molecule 377790), all of which

terial activity have been described (Table 1). These inhibitors are behave as covalent inhibitors [19–21].

categorised into two families according to their mode of action. Six Recently, a new scaffold was found from a whole-cell screen

are known to inhibit DprE1 irreversibly by forming a covalent against Mycobacterium bovis BCG which resulted in the identiﬁca-

adduct with C387 of DprE1 in the same manner as BTZ, whereas tion of benzothiazole-N-oxide (BTO) targeting DprE1 [22]. Unfor-

nine act as competitive noncovalent inhibitors. This review will tunately, several toxicity and mutagenicity issues were associated

focus on the most recent inhibitors and structural studies of DprE1 with this molecule. However, modulating the stereoelectronic

in complex with its inhibitors. properties of the benzothiazole ring in SAR studies led to the

Common characteristics of the covalent inhibitors are the discovery of a novel class of antitubercular agents called cBT [6-

presence of a nitro group and their inactivity against C387A methyl-7-nitro-5-(triﬂuoromethyl)-1,3-benzothiazoles] [22,23].

and C387S DprE1 mutants. PBTZ169 has one of the lowest MICs Although less potent, cBT are non-mutagenic and exhibit an

against M. tuberculosis (0.6 nM) and resulted from a lead optimisa- improved safety proﬁle.

tion campaign [14,18]. PBTZ169 has completed Phase I clinical Genotoxicity is a potential concern for covalent inhibitors

trials and acts in synergy with pyrazinamide and bedaquiline, because nitro-aromatic compounds generally carry a risk of muta-

making this combination a potential new TB drug regimen [18]. genicity [24]; PBTZ169 has been found to be non-mutagenic in

DprE1 has also been identiﬁed as the target of the dinitrobenza- preclinical tests. Two studies have demonstrated that the nitro

mides (DNBs), nitroquinoxalines (with the lead molecule VI-9376) group present on BTZ can be replaced with a pyrrole ring or an

192 www.drugdiscoverytoday.com 527

REVIEWS Drug Discovery Today Volume 22, Number 3 March 2017

azide group while retaining signiﬁcant antimycobacterial activity AQs. This series has excellent drug-like properties, including good

[25,26]. These non-nitro BTZ analogues then behave as noncova- solubility and an attractive pharmacological proﬁle. A new strate-

lent inhibitors and are considerably less potent than their covalent gy to ﬁnd novel inhibitors involves forming hybrids between

counterparts. moieties of two different molecules. Examples include 1,2,4-tria-

Within the past three years, a considerable number of nonco- zole-containing 1,4-benzothiazinone derivatives [35] and a pyra-

valent DprE1 inhibitors have been found. In a cell-based screen, a zine–thiazolidinone hybrid scaffold [36], showing relatively low

new compound, TCA1, was identiﬁed that has activity against activity against M. tuberculosis.

replicating and nonreplicating M. tuberculosis. It is also effective in

Reviews

vivo alone or in combination with frontline TB drugs in acute and Structural studies of DprE1

chronic mouse models of TB. TCA1 targets two different enzymes: There are currently 22 entries in the Protein Data Bank correspond-

OTSCREEN POST DprE1 and MoeW [27]. Lacking a nitro group, it is unable to bind ing to DprE1 structures (Table 2). Three structures of Mycobacterium

covalently to C387. TCA1-resistant mutants harbour a Y314C smegmatis DprE1 have been solved in two different space groups

substitution in DprE1. Interestingly, the Y314C mutant strain, and one structure contains BTZ043. The 19 other structures are M.

which is resistant to TCA1, remains sensitive to BTZ suggesting tuberculosis DprE1 crystallised in two crystal forms, namely hexag-

that the binding mechanism of TCA1 to DprE1 is different from onal and orthorhombic, in complex with or without inhibitors.

that of BTZ. Recently, new molecules based on the structure of The structures of the M. smegmatis and M. tuberculosis DprE1

TCA1, benzothiazolylpyrimidine-5-carboxamides, were designed enzymes are similar (83% sequence identity; RMSD of 0.464 A˚ for

by structure-based drug design approaches. [28]. These new mole- 340 Ca) [16,17,37]. DprE1 is characterised by the two-domain

cules are more active than TCA1 with an MIC of 80 nM (seven-fold topology of the vanillyl-alcohol oxidase family of oxidoreductases

lower than that of TCA1) in M. tuberculosis and have better oral including a FAD-binding domain (residues 7–196, 413–461) and

bioavailability than TCA1 and BTZ043. the substrate-binding domain (residues 197–412). The monoclinic

The 1,4-azaindole series is another class of noncovalent inhi- and hexagonal crystal forms display an apparent dimer of DprE1.

bitors that target DprE1, and was identiﬁed during a scaffold- However, Batt et al. suggest that DprE1 does not dimerise in

morphing approach starting from a published anti-TB, non-DprE1 solution [16].

imidazo-pyridine scaffold [29–31]. Spontaneous resistant mutants The cofactor is deeply buried in the FAD-binding domain, with

contain a single Y314H mutation in DprE1, which is the same the isoalloxazine positioned at the interface of the substrate-

mutation found in TCA1-resistant mutants. In contrast, no cross- binding domain in front of the substrate-binding pocket

resistance was observed between BTZ and azaindole-resistant [16,17]. In contrast to the homologous structure of alditol oxidase,

strains, suggesting that TCA1 and 1,4-azaindoles behave similarly DprE1 does not covalently bind the prosthetic group. Interesting-

to noncovalent inhibitors. Like TCA1, 1,4-azaindoles have good ly, the M. smegmatis DprE1 structure has also been solved in the

oral exposure, are active in vivo and are promising candidates for absence of the FAD cofactor, indicating that FAD is inessential for

the development of a novel anti-TB drug [32]. the folding of the protein. [37].

Pyrazolopyridones, which originated from a whole-cell screen, Electron density in all crystal forms obtained for M. tuberculosis

were also found to inhibit DprE1 in a noncovalent manner with an or M. smegmatis DprE1 reveals disorder of two surface loops in the

MIC of 0.1 mM [33]. As with 1,4-azaindoles, the Y314H mutation substrate-binding domain (Fig. 1) [16–18,27,34,37]. These loops

confers resistance to pyrazolopyridones. Interestingly, pyrazolo- can be involved in interactions with the cell membrane, with other

pyridones showed enhanced potency against the BTZ-resistant protein partners involved in the DPA biosynthesis pathway or with

strains carrying C387S and C387G mutations in DprE1 as com- the substrate DPR [16]. Interestingly, disordered loop I (amino

pared with the wild-type strain. This series has not been tested in acids 269–303) and disordered loop II (316–330) are located just

vivo because the pharmacodynamic properties require further above the substrate-binding pocket, leaving the active site wide

optimisation. open in the ligand-free form and thus could act as a gate for the

2-Carboxyquinoxalines have also been identiﬁed as noncova- substrate. The extent of disorder depends on the contact with the

lent DprE1 inhibitors in a phenotypic screening of a quinoxaline symmetry-related molecule of the crystal form. In all the mono-

library [34]. Spontaneous resistant mutants have G17C and L368P clinic crystal forms (P21) where there are two molecules in the

substitutions in DprE1. Interestingly, although the authors have asymmetric unit, disorder of the two loops is higher in monomer B

indicated that no cross-resistance was observed between either of than in monomer A owing to symmetric contact. Furthermore, it is

these dprE1 mutants and BTZ043, the BTZ-resistant C387S mutant clear from structural studies that the disordered loops can behave

strain conferred some cross-resistance to Ty38c by increasing the differently depending on the inhibitor. Some inhibitors such as

MIC four-fold. These observations suggest that G17, L368 and Ty38c, QN118, QN127 and Ty21c stabilise loop I, whereas CT319

C387 are involved in the binding of 2-carboxyquinoxalines to stabilises loop II. Ty36c and QN129 stabilise both loops (Table 2).

DprE1. Unfortunately, these compounds are inactivated by decar-

boxylation in M. tuberculosis by the protein encoded by rv3406. Structural studies of DprE1 in complex with covalent

The lead compound of 2-carboxyquinoxalines is Ty38c which has inhibitors

an MIC of 3.1 mM. M. smegmatis DprE1 was crystallised in complex with BTZ043 [17],

4-Aminoquinolone piperidine amides (AQs) are novel scaffolds shedding light on the structural basis of the inhibition mecha-

identiﬁed from a whole-cell screen with an MIC of 800 nM [35]. nisms of covalent inhibitors. BTZ043 is a mechanism-based cova-

Again, the Y314H mutation leading to azaindole, pyrazolopyri- lent inhibitor, which requires the enzymatic activity of the protein

done and TCA1 resistance has a strong impact on the efﬁcacy of with the substrate to convert the nitro group of BTZ043 to the

528 www.drugdiscoverytoday.com 193

Drug Discovery Today Volume 22, Number 3 March 2017 REVIEWS

TABLE 2

Solved structure of DprE1 present in PDB

PDB ID Space group Resolution Mol/ASU Disordered regions Species Inhibitor Inhibition Refs

(loop I/loop II) mechanism

4G3T P3221 2.35 1 (275–307)/(323–337) M. smegmatis [37]

4G3U P21 2.69 1 (275–307)/(323–337) M. smegmatis [37]

4FDP P21 2.23 2 Chain A: (269–283)/(316–330) M. tuberculosis [16]

Chain B: (269–293)/(316–330)

4FEH P64 2.04 1 (269–297)/(316–322) M. tuberculosis [16]

4AUT P3221 2.1 1 (275–303)/(330–336) M. smegmatis [17]

4F4Q P212121 2.62 1 (275–302)/(324–337) M. smegmatis BTZ043 Covalent [17]

POST SCREEN

4NCR P21 1.88 2 Chain A: (272–283)/(315–330) M. tuberculosis PBTZ169 Covalent [18]

Chain B: (271–295)/(315–330)

4FDN P64 2.4 1 (269–297)/(315–322) M. tuberculosis CT325 Covalent [16]

Reviews

4FF6 P21 2.6 2 Chain A: (269–283)/(316–330) M. tuberculosis CT325 Covalent [16]

Chain B: (269–292)/(316–330)

4P8H P21 3 2 Chain A: (269–293)/(315–330) M. tuberculosis cBT6a Covalent [23]

Chain B: (269–290)/(314–330)

4PFA P21 2.56 2 Chain A: (271–284)/(317–329) M. tuberculosis BTO Covalent [22]

Chain B: (269–292)/(317–330)

4FDO P64 2.4 1 (269–297)/– M. tuberculosis CT319 Noncovalent [16]

4KW5 P21 2.61 2 Chain A: (269–287)/(317–330) M. tuberculosis TCA1 Noncovalent [27]

Chain B: (268–293)/(317–330)

4P8C P21 1.95 2 Chain A: (269–273)/(324–325) M. tuberculosis QN127 Noncovalent [34]

Chain B: (271–293)/(317–330)

4P8K P21 2.49 2 Chain A: –/(317–325) M. tuberculosis Ty38c Noncovalent [34]

Chain B: (271–292)/(317–330)

4P8L P21 2.02 2 Chain A: –/– M. tuberculosis Ty36c Noncovalent [34]

Chain B: (271–292)/(317–330)

4P8M P21 2.09 2 Chain A: (271–289)/(318–325) M. tuberculosis QN114 Noncovalent [34]

Chain B: (268–300)/(314–329)

4P8N P21 1.79 2 Chain A: –/(327–330) M. tuberculosis QN118 Noncovalent [34]

Chain B: (270–290)/(317–330)

4P8P P21 2.2 2 Chain A: –/(324–329) M. tuberculosis QN127 Noncovalent [34]

Chain B: (271–290)/(317–330)

4P8T P21 2.55 2 Chain A: –/– M. tuberculosis QN129 Noncovalent [34]

Chain B: (271–290)/(316–330)

4P8Y P21 2.01 2 Chain A: –/(327–329) M. tuberculosis Ty21c Noncovalent [34]

Chain B: (271–293)/(315–330)

4PFD P21 2.3 2 Chain A: (268–289)/(320–325) M. tuberculosis cBT Noncovalent [22]

Chain B: (268–298)/(316–329)

Mol/ASU corresponds to the number of molecules per asymmetric unit.

active nitroso form. To generate the DprE1–BTZ043 adduct and at the protein surface and lacks full electron density, resulting in

to obtain the structure of the covalent complex, DprE1 was the ﬂexibility of this region of the molecule. Indeed, there is

incubated with BTZ043 and farnesylphosphoryl-D-ribofuranose only one van der Waals interaction between L363 and the

(FPR; an analogue of DPR with a shorter polyprenyl chain spirocyclic moiety (Fig. 2a). The two disordered regions are

serving as a suitable enzyme substrate) before performing crys- not visible in the density map, suggesting that BTZ043 does

tallisation trials [17]. BTZ043 is located in the substrate-binding not interact with these loops.

pocket in front of the isoalloxazine ring of FAD and binds M. tuberculosis DprE1 was also crystallised in the presence of

covalently to C394 (equivalent to C387 in M. tuberculosis). There PBTZ169 [18], which binds DprE1 in the same manner as BTZ043,

are no major structural changes between the native complexed with the exception of an extra van der Waals interaction with

forms of DprE1. The triﬂuoromethyl group of BTZ043 is situated W230 (Fig. 2b). As with BTZ043, the electron density map does not

in a hydrophobic pocket formed by side chains of H132, G133, account fully for the cyclohexyl moiety of PBTZ169, suggesting

K134, K367, F369 and N385. The piperidine ring of BTZ043 is that this moiety is not stabilised by the protein. Futhermore, loops

maintained on each side by the isoalloxazine ring of FAD, and I and II are also disordered as in the BTZ043–DprE1 complex

by G117 and V365. The spirocyclic moiety of BTZ043 is located suggesting that there are no interactions in these regions.

194 www.drugdiscoverytoday.com 529

REVIEWS Drug Discovery Today Volume 22, Number 3 March 2017

Loop II in complex (a) (b) with Ty36c Loop II in complex Substrate Disordered loop with CT319 Loop II in complex binding pocket II 316 to 330 with QN129 Loop I FAD binding domain

Disordered loop Reviews I 269 to 303 OTSCREEN POST Substrate binding domain

Drug Discovery Today

FIGURE 1

Structure of DprE1 and conformations of the two disordered loops. (a) Overall structure of DprE1 (PDB code 4P8L). FAD-binding domain is in deep blue and

substrate-binding domain in light blue. Substrate-binding pocket is represented as a lime green surface. The two disordered loops are represented in orange for

loop I (residues from 269 to 303) and hot pink for loop II (316 to 330). (b) Superimposition of the three structures where loop II is stabilised in the crystal structure.

Conformations of loop II in the presence of CT329 (violet), QN129 (salmon) or Ty36c (hot pink) are different, suggesting that inhibitors induced the conformational

change and are implicated in the stabilisation of the loop.

Recently, Lange et al. reported the structures of covalent com- part of the molecule is more similar to DNB compounds. The

plexes between DprE1 and BTO derivatives and showed that these compound is located in the BTZ-binding pocket, with the triﬂuor-

compounds also have the same binding mode as BTZ (Fig. 2c). The omethyl group in the hydrophobic groove. Surprisingly, the phenyl

triﬂuoromethyl group occupies the same hydrophobic groove and group of CT319 stabilises the disordered loop II with van der Waals

the covalent bond with C387 remains the main determinant of interactions. However, this complex was made with a nonactivated

interaction (Fig. 2c). DprE1 was crystallised in complex with cBT, enzyme (without prior incubation) and so the nitro group of CT319

which still binds in the same pocket as the original BTO, but does not bind covalently to C387. To circumvent the need for

curiously the methyl group creates an unfavourable interaction inhibitor activation for crystallisation studies, CT325 was synthe-

with neighbouring polar residues (N385, Q336), which affects its sised. It is a nitroso derivative of CT319 that can covalently react

ability to bind covalently to the protein, and could explain the with the cysteine of the active site without the need for prior

decreased activity for this compound (Fig. 2d). A structure of a enzymatic activation. CT325 binds in the same manner as CT319

complex with cBT binding to DprE1 as a noncovalent inhibitor has except that it forms a covalent complex with C387. Interestingly,

been solved (Fig. 2e). Obviously, cBT binds differently to the although the terminal phenylethyl group of CT319 and loop II of

protein compared with covalent inhibitors. The methyl group DprE1 interact and are ordered in the noncovalent complex, they

of cBT is situated in the hydrophobic pocket whereas the triﬂuor- are disordered in the case of the CT325 in the covalent complex.

omethyl group is shifted towards another hydrophobic groove.

The nitro group, which appears not to have been transformed into Structural studies of DprE1 in complex with

a nitroso, is stabilised by H-bonds formed with residues N365 and noncovalent inhibitors

Q336. Finally, C387 and K418 stabilise the complex by the forma- Structural characterisation of TCA1 in complex with M. tuberculosis

tion of additional H-bonds. The structure is not linked to any DprE1 revealed that TCA1 is located in the central cavity of DprE1

publication in the Protein Data Bank and there is no information in a boomerang-like conformation as in the case of the covalent

about the relevance of this complex. Nevertheless, this structure inhibitors, with the thiophene moiety situated deep in the hydro-

could represent the pre-complex before the nitro-reduction of the phobic pocket at the bottom of the active site [27]. As with the

nitro group, suggesting that there is a conformational change of covalent inhibitors, this interaction appears to be very important

the inhibitor before and after nitroreduction by FADH2. Finally, for the binding. Furthermore, TCA1 is maintained by polar con-

loops I and II are not modelled in the three complexes with BTO or tacts between the carboxamido group and thiazole nitrogen of

cBT derivatives, suggesting that there are no interactions between TCA1 and K418 of M. tuberculosis DprE1. Also, the carbamate

these inhibitors and loop II (Table 2). moiety forms van der Waals interactions important for the stabi-

M. tuberculosis DprE1 was also crystallised in complex with lisation of the compound with the phenyl ring of Y314. The

CT319, which is from a class of BTZ-inspired inhibitors [16]. In signiﬁcance of this interaction was proven when Y314 was

this study, complexes were obtained in different space groups, replaced by alanine, leading to a TCA1-resistant mutant [27].

providing crystallographically independent views of the binding Structures with eight different 2-quinoxalines have been solved

mode of inhibitors. CT319 has the nitro-aromatic ring moiety and [34]. Interestingly, disordered loops I and II are visible in the

the triﬂuoromethyl functional group of BTZ043, whereas the other electronic density map to different extents, depending on the

530 www.drugdiscoverytoday.com 195

Drug Discovery Today Volume 22, Number 3 March 2017 REVIEWS

(a) M. smegmatis DprE1 in (b) M. tuberculosis DprE1 in (c) M. tuberculosis DprE1 in (d) M. tuberculosis DprE1 in (e) M. tuberculosis DprE1 in

complex with BTZ043 complex with PBTZ169 complex with BTO complex with cBT37 complex with cBT PDB:4F4Q PDB:4NCR PDB:4PFA PDB:4P8H PDB:4PFD

G117 FAD G117 G117 FAD FAD FAD FAD H132 G117 G117 H132 H132 L363 POST SCREEN L363 H132 H132 G133 W230 G133 L363 G133 G133 F369 F369 W230 G133 W230 F369 K367 K367 F369 F369 K418 K367 K367 C387 K134 C387 K134 C387 C387 K367 K134 K134 Q336 C387 Reviews N385 N385 Q336 K134 N385 N385 N385

(f) M. tuberculosis DprE1 in (g) M. tuberculosis DprE1 in (h) M. tuberculosis DprE1 in (i) M. tuberculosis DprE1 in (j) M. tuberculosis DprE1 in

complex with CT319 complex with CT315 complex with TCAI complex with Ty36c (chain A) complex with Ty36c (chain B)

PDB:4FDO PDB:4FDN PDB:4KW5 W230 PDB:4P8L chain A PDB:4P8L chain B FAD FAD Y314 N324 FAD FAD Loop II FAD H132 H132 H132 Y60 Y60 H132 L317 H132 Y60 L363 G133 G133 L363 G133 G133 G133 F369 F369 F369 F369 G321 Y60 R325 F369 W230 Y60

K367 K367 K367 K418 K367 K418

L317 Loop II K367 K418 W230 K134 K134 K134 K134 K134 N385 C387 C387 C387 C387 N385 N385 C387 N385 N385

Drug Discovery Today

FIGURE 2

Structural and interaction map of DprE1 with inhibitors. Carbon a trace atom is represented in grey and inhibitors by orange sticks. Except for the structure of

Mycobacterium smegmatis DprE1 in complex with BTZ043 (4F4Q), all the complexes were crystallised with the Mycobacterium tuberculosis DprE1 enzyme. To

facilitate description, the M. tuberculosis numbering is used. The hydrophobic groove is represented by sticks and surface is blue. FAD, G117 and V365 (yellow

sticks) form the middle part of the binding cavity and the side of the pocket. In hot pink are amino acids implicated in the stabilisation of inhibitors in the pocket

with van der Waals contacts. Amino acids involved in H-bonding with inhibitors are represented in lime green and the H-bonds by black dots. Cysteine 387 is

represented as a red stick when it forms a covalent bond with the inhibitors and in salmon pink when it is not involved in inhibitor binding.

complex (Table 2). Whereas loop I is complete in chain A of almost might explain the stabilisation of loop II in these two complexes.

all structures, electron density was observed for the whole of loop The two quinoxaline-resistant mutants with G17C and L368P

II in only two cases (chain A with Ty36c and QN129), providing mutations do not directly interact with any of the quinoxaline

important information regarding the ﬂexibility and potential inhibitors, indicating that these mutations could induce a confor-

interactions of this loop with the 2-carboxyquinoxalines. mational change leading to a modiﬁcation of the binding pocket.

In all complexes, the common 2-carboxy-6-triﬂuoromethylqui- Intriguingly, there are two different conﬁgurations of the com-

noxaline core is observed in the same position as with covalent and pounds in the structures. Quinoxaline is able to adopt a planar

noncovalent inhibitors, located close to the FAD. The triﬂuoro- conformation (benzyl and quinoxaline rings approximately copla-

methyl group occupies the same hydrophobic pocket as was ob- nar) as in the Ty36c monomer A complex structure (Fig. 2i) or a bent

served for the covalent inhibitors BTZ043 or PBTZ169. The conformation (benzyl and quinoxaline rings approximately perpen-

carboxylate group of the quinoxaline forms a H-bond with K418, dicular) as in monomer B of the same crystal structure (Fig. 2j).

an essential catalytic residue, and Y60, which is essential for binding

the 2-quinoxalines. Interestingly, in the Ty36c and QN129 com- Concluding remarks and future perspectives

plexes, there is an extra H-bond between the carboxylate group of DprE1 is present only in actinobacteria, and its druggability is

the quinoxaline and the side chain of R325, which is located in linked to its essentiality in mycobacteria because it is required for

disordered loop II. There are also two van der Waals interactions the formation of the mycobacterial cell wall, and to its location in

with L317 and N324 that are involved in stabilisation and these the periplasm. These properties account for the vulnerability of

196 www.drugdiscoverytoday.com 531

REVIEWS Drug Discovery Today Volume 22, Number 3 March 2017

this target. To our knowledge, DprE1 is the only FAD-dependent Waals interactions. In certain cases, amino acids in the substrate-

enzyme belonging to the vanillyl-alcohol oxidase family of oxi- binding pocket form H-bonds with inhibitors, for instance Y60 and

doreductases that possesses a cysteine in the active site, and K418 with the 2-quinoxalines, K418 with TCA1 and N386, Q336

therefore is able to activate and bind nitro-inhibitors. and K418 with the cBT series.

Since the discovery of BTZ043 in 2009, 15 new classes of It is important to note that several docking studies on DprE1

inhibitors were found to be active against M. tuberculosis and target in complex with inhibitors such as benzothiazolylpyrimidine-5-

DprE1. Currently, the most potent of all DprE1 inhibitors, carboxamides [28], 4-aminoquinolone piperidine amides [38],

PBTZ169, is nearing Phase II trials as a result of its high efﬁcacy 1,2,4-triazole-containing 1,4-benzothiazinone derivatives [35],

Reviews

against M. tuberculosis MDR strains and its synergism with BDQ. pyrazolopyridones, 8-pyrrole-benzothiazinones [25] and 1,3-

We believe that this compound could greatly increase the chances benzothiazinone azide [26] have been conducted. The results

OTSCREEN POST of survival of patients affected by multidrug-resistant TB. of these studies were not extensively developed in this review

From the comparative analysis of the 22 structures in PDB, it is but it seems that the same, aforementioned regions contribute

clear that the binding sites of covalent and noncovalent inhibitors to the stabilisation of all the inhibitors. Up until now, the

overlap signiﬁcantly and all the inhibitors target the same binding majority of the inhibitory scaffolds have arisen from whole-cell

pocket. The interaction of a hydrophobic group with the hydro- screening campaigns. Nowadays, attempts are underway to use

phobic groove is common to all the inhibitors. This interaction more-rational design to increase compound activity using the

appears to be the ﬁrst key determinant for binding. C387 clearly available structural information, not only to guide drug design

plays a crucial part for the binding of covalent inhibitors, and also and SAR but also to improve pharmacodynamic properties and

remains an important residue in the binding of noncovalent inhi- bioavailability. Our review will aid the design of such new

bitors such as 2-quinoxalines and pyrazolopyridones. Nevertheless, inhibitors.

the covalent link with DprE1 might be dispensable for the binding

as was observed for CT319 or cBT, which could bind to DprE1 Conﬂicts of interest

without forming a covalent bond. Interestingly, some inhibitors S.T.C. is a named inventor on patents related to this article.

such as Ty36c, CT319 or QN129 are stabilised by the disordered

loop II. Altogether, these structures provide interesting information Acknowledgement

on potential contacts such as R325 with the disordered loop II, This work was supported by the European Community’s Seventh

which could be exploited to enhance the afﬁnity of new inhibitors Framework Programme FP7/2007–2013 under grant agreement

for DprE1. Inhibitors are mainly stabilised in the pocket by van der 260872.

References

1 Kieser, K.J. and Rubin, E.J. (2014) How sisters grow apart: mycobacterial growth and 15 Trefzer, C. et al. (2012) Benzothiazinones are suicide inhibitors of mycobacterial

division. Nat. Rev. Microbiol. 12, 550–562 decaprenylphosphoryl-beta-D-ribofuranose 2 -oxidase DprE1. J. Am. Chem. Soc. 134,

2 Takayama, K. and Kilburn, J.O. (1989) Inhibition of synthesis of arabinogalactan by 912–915

ethambutol in Mycobacterium smegmatis. Antimicrob. Agents Chemother. 33, 1493–1499 16 Batt, S.M. et al. (2012) Structural basis of inhibition of Mycobacterium tuberculosis

3 Quemard, A. et al. (1991) Isoniazid inhibition of mycolic acid synthesis by cell DprE1 by benzothiazinone inhibitors. Proc. Natl. Acad. Sci. U. S. A. 109, 11354–11359

extracts of sensitive and resistant strains of Mycobacterium aurum. Antimicrob. Agents 17 Neres, J. et al. (2012) Structural basis for benzothiazinone-mediated killing of

Chemother. 35, 1035–1039 Mycobacterium tuberculosis. Sci. Transl. Med. 4, 150ra121

4 Mikusova, K. et al. (1995) Biogenesis of the mycobacterial cell wall and the site of 18 Makarov, V. et al. (2014) Towards a new combination therapy for tuberculosis with

action of ethambutol. Antimicrob. Agents Chemother. 39, 2484–2489 next generation benzothiazinones. EMBO Mol. Med. 6, 372–383

5 Belanger, A.E. et al. (1996) The embAB genes of Mycobacterium avium encode an 19 Christophe, T. et al. (2009) High content screening identiﬁes decaprenyl-

arabinosyl transferase involved in cell wall arabinan biosynthesis that is the target for phosphoribose 2 epimerase as a target for intracellular antimycobacterial

the antimycobacterial drug ethambutol. Proc. Natl. Acad. Sci. U. S. A. 93, 11919–11924 inhibitors. PLoS Pathog. 5, e1000645

6 Vilcheze, C. et al. (2006) Transfer of a point mutation in Mycobacterium tuberculosis 20 Magnet, S. et al. (2010) Leads for antitubercular compounds from kinase inhibitor

inhA resolves the target of isoniazid. Nat. Med. 12, 1027–1029 library screens. Tuberculosis 90, 354–360

7 Li, W. et al. (2016) Therapeutic potential of the Mycobacterium tuberculosis mycolic 21 Stanley, S.A. et al. (2012) Identiﬁcation of novel inhibitors of M. tuberculosis

acid transporter, MmpL3. Antimicrob. Agents Chemother. 60, 5198–5207 growth using whole cell based high-throughput screening. ACS Chem. Biol. 7,

8 Riccardi, G. et al. (2013) The DprE1 enzyme, one of the most vulnerable targets of 1377–1384

Mycobacterium tuberculosis. Appl. Microbiol. Biotechnol. 97, 8841–8848 22 Landge, S. et al. (2015) Discovery of benzothiazoles as antimycobacterial agents:

9 Wolucka, B.A. (2008) Biosynthesis of D-arabinose in mycobacteria – a novel bacterial synthesis, structure–activity relationships and binding studies with Mycobacterium

pathway with implications for antimycobacterial therapy. FEBS J. 275, 2691–2711 tuberculosis decaprenylphosphoryl-beta-D-ribose 2 -oxidase. Bioorg. Med. Chem. 23,

10 Mikusova, K. et al. (2005) Decaprenylphosphoryl arabinofuranose, the donor of the 7694–7710

D-arabinofuranosyl residues of mycobacterial arabinan, is formed via a two-step 23 Landge, S. et al. (2016) Nitroarenes as antitubercular agents: stereoelectronic

epimerization of decaprenylphosphoryl ribose. J. Bacteriol. 187, 8020–8025 modulation to mitigate mutagenicity. ChemMedChem 11, 331–339

11 Crellin, P.K. et al. (2011) Decaprenylphosphoryl-beta-D-ribose 2 -epimerase, the 24 Purohit, V. and Basu, A.K. (2000) Mutagenicity of nitroaromatic compounds. Chem.

target of benzothiazinones and dinitrobenzamides, is an essential enzyme in Res. Toxicol. 13, 673–692

Mycobacterium smegmatis. PLoS ONE 6, e16869 25 Makarov, V. et al. (2015) The 8-pyrrole-benzothiazinones are noncovalent

12 Kolly, G.S. et al. (2014) Assessing the essentiality of the decaprenyl-phospho-D- inhibitors of DprE1 from Mycobacterium tuberculosis. Antimicrob. Agents Chemother.

arabinofuranose pathway in Mycobacterium tuberculosis using conditional mutants. 59, 4446–4452

Mol. Microbiol. 92, 194–211 26 Tiwari, R. et al. (2016) Design, syntheses, and anti-TB activity of 1,3-

13 Brecik, M. et al. (2015) DprE1 is a vulnerable tuberculosis drug target due to its cell benzothiazinone azide and click chemistry products inspired by BTZ043. ACS Med.

wall localization. ACS Chem. Biol. 10, 1631–1636 Chem. Lett. 7, 266–270

14 Makarov, V. et al. (2009) Benzothiazinones kill Mycobacterium tuberculosis by 27 Wang, F. et al. (2013) Identiﬁcation of a small molecule with activity against drug-

blocking arabinan synthesis. Science 324, 801–804 resistant and persistent tuberculosis. Proc. Natl. Acad. Sci. U. S. A. 110, E2510–E2517

532 www.drugdiscoverytoday.com 197

Drug Discovery Today Volume 22, Number 3 March 2017 REVIEWS

28 Chikhale, R. et al. (2015) Development of selective DprE1 inhibitors: design, 34 Neres, J. et al. (2015) 2-Carboxyquinoxalines kill Mycobacterium tuberculosis through

synthesis, crystal structure and antitubercular activity of benzothiazolylpyrimidine- noncovalent inhibition of DprE1. ACS Chem. Biol. 10, 705–714

5-carboxamides. Eur. J. Med. Chem. 96, 30–46 35 Shaikh, M.H. et al. (2016) Synthesis and bioactivity of novel triazole incorporated

29 No, Z. et al. (2011) Anti-infective compounds. WO 2011/113606 A1 benzothiazinone derivatives as antitubercular and antioxidant agent. Bioorg. Med.

30 No, Z. et al. (2011) Anti-infective pyrido(1,2-a)pyrimidines. WO 2011/085990 A1 Chem. Lett. 26, 561–569

31 Shirude, P.S. et al. (2013) Azaindoles: noncovalent DprE1 inhibitors from scaffold 36 Chitre, T.S. et al. (2016) Synthesis and docking studies of pyrazine-thiazolidinone

morphing efforts, kill Mycobacterium tuberculosis and are efﬁcacious in vivo. J. Med. hybrid scaffold targeting dormant tuberculosis. Bioorg. Med. Chem. Lett. 26,

Chem. 56, 9701–9708 2224–2228

32 Chatterji, M. et al. (2014) 1,4-Azaindole, a potential drug candidate for treatment of 37 Li, H. and Jogl, G. (2013) Crystal structure of decaprenylphosphoryl-beta-D-ribose

tuberculosis. Antimicrob. Agents Chemother. 58, 5325–5331 2 -epimerase from Mycobacterium smegmatis. Proteins 81, 538–543

33 Panda, M. et al. (2014) Discovery of pyrazolopyridones as a novel class of 38 Naik, M. et al. (2014) 4-Aminoquinolone piperidine amides: noncovalent inhibitors

noncovalent DprE1 inhibitor with potent anti-mycobacterial activity. J. Med. Chem. of DprE1 with long residence time and potent antimycobacterial activity. J. Med.

57, 4761–4771 Chem. 57, 5419–5434 POST SCREEN Reviews

198 www.drugdiscoverytoday.com 533 Appendix Chapter A2: Structure-based drug design and characterization of sulfonyl-piperazine benzothiazinone inhibitors of DprE1 from Mycobacterium tuberculosis

Jérémie Piton1, Anthony Vocat1, Andréanne Lupien1, Caroline Foo1, Olga Riabova2, Vadim Makarov2, Stewart T. Cole1*

1Global Health Institute, École Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland 2FRC Fundamentals of Biotechnology RAS, Leninsky prospekt 33-2, Moscow, 119071, Russian Federation

Antimicrobial Agents and Chemotherapy, 2018, doi: 10.1128/AAC.00681-18

Contributions: execution of microsomal and enzymatic assays, manuscript preparation

199

ABSTRACT

Macozinone (MCZ) is a tuberculosis (TB) drug candidate that specifically targets the essential flavoenzyme DprE1 thereby blocking synthesis of the cell wall precursor decaprenyl phosphoarabinose (DPA) and provoking lysis of Mycobacterium tuberculosis. As part of the MCZ back-up program we exploited structure-guided drug design to produce a new series of sulfone- containing derivatives, 2-sulphonylpiperazin 8-nitro 6-trifluoromethyl 1,3-benzothiazin-4-one, or sPBTZ. These compounds are less active than MCZ but have a better solubility profile and some derivatives display enhanced stability in microsomal assays. DprE1 was efficiently inhibited by sPBTZ and covalent adducts with the active site cysteine residue (C387) were formed. However, despite the H-bonding potential of the sulfone group no additional bonds were seen in the crystal structure of the sPBTZ-DprE1 complex with compound 11326127 as compared to MCZ.

Compound 11626091, the most advanced sPBTZ, displayed good antitubercular activity in the murine model of chronic TB but was less effective than MCZ. Nonetheless, further testing of this

MCZ backup compound is warranted as part of combination treatment with other TB drugs.

201 INTRODUCTION

DprE1 is an essential flavoprotein of Mycobacterium tuberculosis involved in decaprenylphosphoryl-beta-D-arabinose (DPA) synthesis. DPA is the sole precursor of arabinose for production of both arabinogalactan and lipoarabinomannan (1), important components of the mycobacterial cell wall. DprE1, together with its counterpart DprE2, catalyses the epimerization of decaprenylphosphoryl-ß-D-ribose (DPR) to DPA in a two-step mechanism. In the last decade, many inhibitors were discovered to target DprE1, which is considered nowadays as the Achilles’ heel of M. tuberculosis due to its essentiality and especially to its localization in the periplasm (2).

DprE1 inhibitors can be classified into two families according to their mode of action: some of them inhibit DprE1 irreversibly by forming a covalent adduct with cysteine 387 (C387) of DprE1, whereas others act as competitive non-covalent inhibitors (3).

The first covalent DprE1 inhibitors discovered were benzothiazinones (BTZ), exemplified by the lead compound, BTZ043, which is exceptionally potent with in vitro and ex vivo minimal inhibitory concentration (MIC) values in the nanomolar range (4). A lead optimization campaign gave rise to PBTZ169, now known as Macozinone (MCZ). It is currently the most potent BTZ compound against M. tuberculosis with an MIC of 0.3 nM (5), has completed preclinical development successfully and is now undergoing Phase I and phase II clinical trials

(https://www.newtbdrugs.org/pipeline/clinical). A common characteristic of the covalent DprE1 inhibitors is the presence of a nitro group on the molecule, which is essential for the mechanism of inhibition. Indeed, this nitro group is converted by DprE1 containing FADH2 into an extremely reactive nitroso group which specifically targets the cysteine residue at position 387 (C387) in the active site of DprE1, to form a covalent adduct and thereby irreversibly inhibits the enzyme (4-7,

8 , 9-13). It has been demonstrated that the presence of C387 is essential for the activity of covalent

DprE1 inhibitors (4, 14).

202 Apart from the covalent bond with C387, covalent DprE1 inhibitors are otherwise only maintained in the pocket by steric hindrance and Van Der Waals interactions, which explains why a simple substitution at C387 results in complete resistance of the enzyme to these compounds (3).

As part of the MCZ-back-up program, this observation prompted us to revisit the Structure Activity

Relationship (SAR) in order to obtain derivatives with other anchor points in the active site of

DprE1, since such compounds might retain activity against C387 mutants should these arise.

Several non-covalent DprE1 inhibitors have also been found (15-24). Similar to the covalent DprE1 inhibitors, these non-covalent compounds sit in the substrate-binding pocket of

DprE1 and act as competitive inhibitors. Interestingly, a class of non-covalent inhibitors, 2- carboxyquinoxalines, are active against BTZ-resistant M. tuberculosis strains with substitutions at

C387 of DprE1 (14). Molecules from this family possess an essential 2-carboxylate moiety that forms key hydrogen bonds with the side-chain of Lysine 418 and the hydroxyl group of Tyrosine

60 (18). Hence, it can be hypothesised that a composite molecule between 2-carboxyquinoxalines and MCZ could overcome resistance issues and increase specificity to the target.

Revisiting the SAR of MCZ further provides the opportunity to improve pharmacodynamics properties of MCZ such as aqueous solubility to increase its oral bioavailability, metabolic stability, and in vivo activity (25). It has been observed that the potency of BTZ derivatives is inversely proportional to their solubility (4, 5). Therefore, the solubility and bioavailability of the most active benzothiazinones are parameters for improvement. Factors controlling the aqueous solubility of organic molecules are complex and drug solubility issues are usually solved by a combination of empirical and rational drug design strategies. More than twenty crystal structures of DprE1 with or without inhibitors were reviewed recently to identify the structural determinants for activity and to guide rational drug design (3).

203 Based on our prior observations, the aim of this study was to design a new structure-guided series of MCZ derivatives with increased activity against either wild-type M. tuberculosis or its

BTZ-resistant mutants, and with improved solubility, absorption, bioavailability and metabolic stability of the compound in vivo. Therefore, a new series of MCZ derivatives, harbouring a sulfonylpiperazine group, was designed (2-sulphonylpiperazin 8-nitro 6-trifluoromethyl 1,3- benzothiazin-4-one or sPBTZ), synthesized and characterized. This study identifies 11626091 as the best sPBTZ and demonstrates that this compound has a promising combination of antitubercular activity and ADME/T (Absorption, Distribution, Metabolism, and Excretion –

Toxicity) properties.

RESULTS

Rationale. When BTZ inhibitors bind to their target, DprE1, the sole bond formed is a covalent semimercaptal bond with the active site cysteine residue, C387. We reasoned first that introducing a sulfonyl group into the MCZ scaffold might offer a second anchor by mimicking the carboxylate moiety of the 2-carboxyquinoxalines that acts as a H-bond acceptor with DprE1 and increases affinity of the inhibitor for the target. Secondly, sulfonyl groups are well characterized and present in many FDA-approved drugs, in particular anti-mycobacterial agents, such as dapsone. Based on these observations, sulfonyl groups might increase both the solubility and bioavailability of the inhibitor in vivo. Lastly, the geometry imposed by sulfonyl groups opens new directions for investigation of the SAR.

Synthesis of sPBTZ. The synthesis of 17 sulfanyl-piperazino BTZ (sPBTZ) derivatives was performed in a two-step procedure from 2-(methylthio)-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one, as described previously (5). Its reaction with a 5-molar excess of free piperazine generated the corresponding piperazine derivative with a high yield. This scaffold was

204 used in the reactions with different alkyl-, aryl- or heteryl-sulfochlorides to form sulfanyl- piperazino BTZs. The compounds synthesized have different types of sulfonyl substitutions thus allowing the structure-activity relationship to be studied. It is clear that compounds with alkyl substitutions have much better antitubercular activity and aryl derivatives have much lower activity, consistent with our previous data for piperazine-containing BTZ (PBTZ) derivatives (5)

Hydrophilicity. The octanol-water partition coefficient, logP, which is regarded as a suitable indicator of molecular hydrophobicity and bioavailability, was calculated for all derivatives to measure the effects of introducing a sulfonyl group on the PBTZ backbone.

Interestingly, the addition of the sulfonyl group between the benzothiazinone and piperazine moieties has the tendency to decrease the clogP coefficient and therefore hydrophobicity (Table 1).

The sulfonylated derivative of MCZ, sPBTZ169 (11326127) that carries a sulfonyl group between the piperazine and cyclohexyl moieties, has a clogP of 3.28 whereas MCZ has a clogP of 4.31. This implies that the introduction of a sulfonyl group decreases hydrophobicity, and may thus increase solubility in physiological conditions and subsequently could have an important impact on bioavailability.

Antitubercular activity. The activity of all sPBTZ derivatives was tested in vitro against

M. tuberculosis H37Rv and MIC99 values were determined (Table 1). Most sPBTZ were active in vitro in the sub-micromolar range, proving that that addition of the sulfonyl group does not abolish activity. However, none has a better activity than MCZ. 11626093, which has a butyl group, has the highest activity of MIC 1 ng/μL corresponding to 5 times the MIC99 of MCZ. The introduction of a sulfone between the piperazine and the cyclohexyl negatively affected the activity of the compound, reducing activity by 30 fold, as observed with sPBTZ169 (11326127), which has an

MIC99 of 6 nM. It was previously shown that substituting the methylcyclohexyl in MCZ with small radicals such as methyl or ethyl decreases the activity of compounds, with MIC of 250 ng/mL and

205 60 ng/mL, respectively (5). Interestingly, when methylcyclohexyl is substituted by sulfonylmethyl

(11626095) or sulfonylethyl (11626092), the activity decreases less and the compounds are 10- times more active than the non-sulfonated derivatives (MIC of 20 ng/mL and 3 ng/mL, respectively). Substitution with a butyl leads to the same activity independent of the sulfonyl group.

This observation indicates that the presence of the sulfonyl group positively affects activity when the radical is small (methyl or ethyl) whereas it negatively influences activity when the substituent is long and hydrophobic.

Similarly to the other BTZ derivatives, the activities of the sulfonyl derivatives are inversely proportional to the logP (4), suggesting that activity could be related to solubility in lipids (Figure

S1). Seven molecules were selected based on their activity/logP profile for further characterisation

(Figure S1).

Target engagement and structural studies. To ensure that sPBTZs specifically target

DprE1, selected sulfonyl derivatives were tested against M. tuberculosis NTB1, a DprE1 mutant that carries a cysteine 387 serine (C387S) substitution and is thus resistant to BTZ. As expected, none of the sBTZs are active against NTB1, indicating that DprE1 is their primary target (Table

2). Furthermore, it is unlikely that the sulfonyl group is able to mimic the carboxylate moiety of 2- carboxyquinoxalines in stabilizing the compound in the pocket as a non-covalent inhibitor.

To investigate whether the introduction of the sulfonyl group could influence the position of the inhibitor and allow more contact within the binding pocket, a crystal structure of DprE1 in complex with sPBTZ169 was solved and compared to the crystal structure of DprE1 with MCZ. sPBTZ169 is located in the same pocket as MCZ and other BTZ derivatives. It sits in an hydrophobic pocket via the trifluoromethyl group and is covalently bound to C387 (Figure 1). The compound is maintained by Van der Waals interaction on each side by V365, Y314, W230 and

FAD, and a hydrogen bond between K418 and the oxygen atom of the nitro group of sPBTZ169.

206 Unfortunately, the orientation of sPBTZ does not favor the formation of a new anchor to the protein for instance interaction between the sulfonyl function and Y60. Futhermore as MCZ, the electron density map does not account fully for the sulfonyl-cyclohexyl moiety of sPBTZ169 likely due to its higher flexibility demonstrating that sulfonyl is not stabilized in the pocket (5).

To determine whether the sulfonyl group could affect the activity at the protein level and even help to stabilize the inhibitor in the active site of a BTZ-resistant C387S DprE1 variant, the inhibition of DprE1 activity was measured in vitro.IC50 values were determined for the wild type and the BTZ-resistant C387S mutant enzyme for MCZ and sPBTZ169 (Table S2). sPBTZ169 has an IC50 of 1.1 and 12 μM against wild type and BTZ-resistant DprE1, respectively, whereas the corresponding IC50 values for MCZ are 0.3 and 3.6 μM. That introducing the sulfonyl group on the

MCZ scaffold leads to 4-fold less activity suggests that even if the environment of the protein is favorable for an H-bond acceptor, the presence of a hydrophobic group, such as a cyclohexyl in

MCZ, is still preferable for activity of the drug. On the other hand, the higher IC50 against C387S compared to the WT enzyme validates the structural studies in that the sulfonyl group does not help to stabilize the molecule in the pocket of the BTZ-resistant mutant C387S.

ADME/T. To assess potential cytotoxic effects of the sulfonyl group on the sPBTZ derivatives, viability of HepG2 cells was monitored after exposure to different concentrations of the compounds. The concentration for half-maximal cytotoxicity (TD50) was determined for each compound. Four of the seven compounds were not cytotoxic (11326059, 11326127, 11626093,

11626094) while three of them showed mild cytotoxicity at concentrations of around 10 μg/mL

(11626091, 11626092, 11626095). Taken together, the selective index representing the ratio of the antitubercular activity of compounds (MIC99 against H37Rv) to cytotoxicity (TD50 against HepG2 cells) is more than acceptable for the chosen seven compounds (Table 2).

207 Since solubility issues are often encountered in drug development, which would consequently impact bioavailability, activity in vitro, ADMET results and activity in vivo, the solubility of these derivatives was calculated by the shake flask method in equilibrium in water.

Experimental solubility constants measured were then compared to theoretical solubility constants calculated using the SwissADME webserver (26). sPBTZ are predicted to have increased solubility compared to MCZ in water (Table 3) and experimental solubility constants measured in water for all sPBTZ derivatives are consistent with theoretical constants calculated by SwissADME using different algorithms (Table 3). As expected, 11626095, 11626092 and 11626091, harboring methyl, ethyl and cyclopropyl groups, respectively, are the most soluble sPBTZ derivatives. In contrast, 11326059, 11326127, 11626093, and 11626094 that carry bigger hydrophobic groups are at least 100-times less soluble in water. Interestingly, there is a good correlation between water solubility and clogP. However, a discrepancy between the predicted and experimental solubility for MCZ was observed. In fact, MCZ is 400 times more soluble in water than calculated with different algorithms.

Another possible issue in the development of MCZ might be metabolic instability (5). In order to test if the sulfonyl group could improve metabolic stability effected by mouse CD-1 or human liver enzymes, microsomal stability experiments were conducted. Intrinsic clearance (Clint) values were then calculated for each compound and compared to the Clint of Carbamazepine and

Nifedipine, used as controls for low and high intrinsic clearance respectively. It is important to note that the Compounds 11626095, 11626091 and 11626092 harbouring the smallest radical chains methyl, ethyl and cyclopropyl, respectively, have the lowest clearances in both mouse and human microsomes, indicative of their stability. 11626093 carrying a butyl radical is metabolically stable in human but highly unstable in mouse microsomes, suggesting that it could be a good substrate for mouse but not human microsomal enzymes. Compounds 11326059 and 11326127 have

208 medium clearances in both mouse and human microsomes, similar to MCZ and BTZ043. Of note,

11626094 is highly unstable in the presence of human microsomes. These results reveal a strong correlation between the length of the substituent after the piperazine moiety and instability in microsomes.

Activity in the murine model of chronic TB. Finally, the in vivo efficacy of 11626091 was assessed in the murine model of chronic TB following low‐dose aerosol infection of

BALB/cByJ mice with M. tuberculosis and treatment at 50 mg/kg (Figure 2). Compound 11626091 was selected as it is the most promising taking into account its activity (“only” 15 times less active than MCZ), cytotoxicity, solubility, and metabolic stability properties. In vivo activity of 11626091 was tested at 50 mg/kg and compared to activity of MCZ at standard dose of 25 mg/kg and INH at standard dose of 10 mg/kg.

Compared to the untreated control group, the bacterial burden in the lungs and spleens of

11626091‐treated mice was 0.46 (P<0.05) and 1.03 (P<0.0001) log10 CFU lower, respectively (Fig

2), whereas the values for those organs from MCZ-treated mice were 1.03 (P<0.0001) and 1.47

(P<0.0001) log10 CFU. These results indicate that 11626091 is highly active in vivo in the murine model of chronic TB, although it is less able to reduce the bacterial load in the lungs than treatment with MCZ.

DISCUSSION

Lead compounds for new anti–TB drugs should possess not only potent sterilizing activity but also good pharmacokinetic profiles to facilitate their co-administration with other anti-TB and anti-HIV

(human immunodeficiency virus) agents. New drugs should also be appropriate for once daily oral dosing and be relatively inexpensive to produce to ensure that all high-burden countries have

209 access. Structure-based rational drug design supports drug development and has the potential to increase the activity and pharmacokinetic properties of lead compounds.

The new sPBTZ series was designed as part of the MCZ back-up program. MCZ is a BCS

(biopharmaceutical classification system) class 2 drug with a low dissolution rate but excellent absorption (5). Structural studies indicate that the environment of the binding pocket in the protein can accommodate a polar group at the cyclohexyl position of MCZ (3). A sulfonyl group was deemed to be a good candidate because it is well characterized and present in many soluble and metabolically stable FDA-approved drugs including those for treating mycobacterial infections.

For instance, sulfamethoxazole is a sulfonamide drug used in co-trimoxazole prophylaxis for HIV- infected patients, demonstrating that it is compatible with anti-retroviral treatment (27).

It was previously shown that there is a strong correlation between logP and BTZ activity

(4). We found that introduction of a sulfonyl group into the MCZ scaffold to form the sPBTZ series decreases both logP and activity in vitro (Table 1). However, sPBTZ still retain potency and the presence of the sulfonyl group improved aqueous solubility for those derivatives which harbour small side chains, such as methyl, ethyl or cyclopropyl, as compared to MCZ (5). Interestingly, aqueous solubility and metabolic stability measured in microsomes indicate a better profile for sPBTZ with small groups rather than those with long hydrophobic chains or MCZ. Despite being less active than MCZ our sulfonylated PBTZ with long hydrophobic chains, methyl (11626095), ethyl (11626092) or cyclopropyl (11626091) derivatives are considered as good candidates in terms of their solubility and metabolic stability profiles. The efficacy of the most active derivative of the three, 11626091, was assessed in vivo in the murine model of chronic TB, where relatively good activity was measured in the lungs and particularly in the spleens (Figure 2). It is also important to note that dissolution of 11626091 was considerably easier than MCZ in methylcellulose, the solvent used for the in vivo studies (technical observation, data not shown).

210 One of the objectives of inserting a sulfonyl group into the PBTZ backbone was to mitigate potential BTZ-resistance by increasing the number of hydrogen bonds with DprE1. Indeed, a sulfonyl could act as an H-bond acceptor in order to anchor the protein to the H-bond donors, for example the hydroxyl group of Tyrosine 60 localized in the binding pocket. This group was identified as a key player in the stabilization of 2-carboxyquinoxalines, molecules that remain active against BTZ-resistant DprE1. Unfortunately, as was observed in the crystal structure and enzymatic inhibition assays, the sulfonyl group of sPBTZ is not implicated in the binding and stabilization of the drug in the pocket as originally hypothesized. This explains the resistance of the BTZ-resistant M. tuberculosis mutant NTB1 to sPBTZ.

To conclude, our study identifies 11626091 as an active, metabolically stable and moderately soluble molecule that is less active than MCZ in vitro and in vivo. However, given its better solubility, compound 11626091 represents an attractive back-up to MCZ, that should now be tested in combination with other TB drug candidates such as bedaquiline in order to assess its potential to contribute to a new regimen.

211 MATERIALS AND METHODS

Synthesis. The synthetic route used to produce sPBTZ and related procedures are described in the supporting materials.

Octanol-water partition coefficient logP calculation. The logP values were calculated using the program Hyperchem 7.5 (Hypercube Inc., http://www.hyper.com).

Bacterial strains and culture conditions. M. tuberculosis strain H37Rv and its BTZ- resistant mutant (C387S) NTB1 were grown at 37°C with shaking in Middlebrook 7H9 broth

(Difco) supplemented with 10% albumin-dextrose-catalase (ADC) enrichment, 0.2% glycerol, and

0.05% Tween 80. The in vitro activities against all mycobacterial strains were measured with the resazurin reduction microplate assay (REMA), by 2-fold serial dilutions of the compounds in the working bacterial culture in 96-well plates (final volume of 100 μl). Following incubation for 6 days at 37°C bacterial viability was determined by adding resazurin for 24 h at 37°C and measuring the fluorescence of the resorufin metabolite (excitation wavelength, 560 nm; emission wavelength,

590 nm) using a Tecan Infinite M200 microplate reader. Briefly, the noise signal obtained in the blank control wells was subtracted from the fluorescence values of test samples and mycobacterial viability in each well were proportionally calculated compared to 100% growth in control wells.

Bacterial viability curves and MIC99 value were calculated with GraphPad Prism software version

7.0, using “Gompertz equation for MIC determination” analysis (GraphPad Software, Inc., La

Jolla, CA, USA).

Cytotoxicity studies. The cytotoxicity of the compounds was measured as described previously against the human hepatic cell line, HepG2 (20). Briefly, cells were incubated (4,000 cells/well) with serial dilutions of compounds (2-fold dilutions; 100 to 0.1 μg/ml) in a 96-well microplate. Following incubation for 3 days at 37°C cell viability was determined by adding resazurin for 4 h at 37°C and measuring the fluorescence of the resorufin metabolite (excitation

212 wavelength, 560 nm; emission wavelength, 590 nm) using a Tecan Infinite M200 microplate reader. Data were corrected for background (no-cell control) and expressed as a percentage of the value for untreated cells (cells only).

Data were fitted to obtain IC50 values, using “Log(inhibitor) vs. response – Variable slope” function implemented in GraphPad Prism software version 7.0 (GraphPad Software, Inc., La Jolla,

CA, USA). Selective index refers to the ratio of the dose of drug that causes toxicity effects (TD50) to the dose that leads to the desired pharmacological effect (MIC99).

Water Solubility. Water solubility measurements were performed using the shake flask method (28). Approximately 1 mL of water was added to 1 mg of compound in an Eppendorf tube and incubated for 3 days at 25°C with shaking at 800 rpm. Suspensions were centrifuged at 16,100 x g for 10 min and the supernatant filtered using 0.22 μ filters. Filtrates were injected onto a high- performance liquid chromatography (HPLC) column (Dionex) and the amount of compound quantified using a calibration curve.

Inhibition assays, crystallography and structural studies of DprE1 complexed with sPBTZ. Recombinant M. tuberculosis DprE1 was overexpressed and purified as described previously (5) to obtain highly concentrated and pure protein with bound flavin adenine dinucleotide (FAD). Enzyme activities in the presence of MCZ, Ty38c and 11326127 were measured as described previously to determine IC50 (50% inhibitory concentrations) values for both wild-type and C131S mutant DprE1 proteins (14).

For crystallization purposes, M. tuberculosis DprE1 (approximately 40 ȣM) was incubated for 3h at 30 °C, with 200 ȣM sulfonyl-BTZ 11326127 and 200 ȣM FPR (farnesyl phosphoribose), in

20 mM Tris pH 8.5, 50 mM NaCl. The protein was concentrated to approximately 15 mg/mL on an Amicon centrifugal device (30,000 MWCO, Millipore). Crystals were obtained by the hanging-

213 drop vapor diffusion method at 18 °C. Experiments were set up by mixing 1 ȣl of the protein sample with 1 ȣl of the reservoir solution containing 100 mM imidazole, pH 7.2-7.5, 18-24 % polypropyleneglycol 400. Yellow crystals grew in approximately 1-3 days and were transferred to a cryo-protectant (reservoir solution with 25% glycerol) prior to flash-cooling in liquid nitrogen.

X-ray data were collected at SLS, beamline PROXIMA 3. Data were integrated with the program XDS (29) and processed using PHENIX (30). Structure was solved by molecular replacement using PHENIX (31) and the structure of M. tuberculosis DprE1 (PDB code : 4NCR,

(5)) as a template. Molecular replacement was subjected to iterative rounds of refinement and rebuilding in coot (32) and PHENIX. The coordinates and structure factors have been deposited in the Brookhaven Protein Data Bank (accession numbers 6G83).

Metabolic stability in vitro. The intrinsic clearance (CLint) of compounds was measured in both mouse and human liver microsomes. Briefly, 100 μg of mouse (CD-1) or human liver microsomes (both from Invitrogen) were mixed in 0.1M phosphate buffer (pH 7.4) containing 0.01

μl of compound dissolved in DMSO at 10 mg/ml, in a final volume of 50 μl. In parallel, a NADPH- regenerating system (Promega) was prepared in 0.1 M phosphate buffer (pH 7.4). The solutions were pre-incubated at 37°C for 10 min before the intrinsic clearance assessment was initiated by mixing the two solutions (50 μl of each; final compound concentration, 1 μg/ml) at 37°C. After 0,

5, 15, 30, and 60 min, the reactions were terminated by transferring 100 μl of the reaction mixture into 100 μl of acetonitrile and placing the mixture on ice for 30 min for full protein precipitation.

Samples were then centrifuged at 12,000 x g for 10 min, and the supernatant was injected onto a high-performance liquid chromatography (HPLC) column (Dionex) to quantify the amount of parent compound remaining over time. To control that the clearance is due to the enzyme activity of the microsomes and not to unspecific microsomal protein binding, mouse (CD-1) or human liver

214 microsomes (both from Invitrogen) were mixed in 0.1M phosphate buffer (pH 7.4) containing 0.01

μl of compound dissolved in DMSO at 10 mg/ml and incubated without NADPH-regenerating system. After 0 and 60 min the reactions were terminated, prepared and analyzed as previously described. Carbamazepine and Nifedipine at the same concentrations were used as controls for low and high intrinsic clearances respectively.

Anti-mycobacterial activity of 11626091 against chronic TB in mice. Female

BALB/cByJ mice, aged 5 to 6 weeks, were purchased from Charles River Laboratories (France).

Mice were infected with a low dose aerosol (100 – 200 CFU/lung) of logarithmic-phase M. tuberculosis H37Rv bacilli, then were allocated to experimental groups and returned to their cages.

Five mice were used per time point for each regimen. Treatment was initiated 4 weeks after infection.

Macozinone (MCZ), 11626091, and isoniazid (INH) were prepared weekly in 0.5% methylcellulose and administered at 25, 50, and 10 mg/kg, respectively, by gavage 5 days a week for 4 weeks. In vivo efficacy of each treatment was assessed by CFU enumeration after plating dilutions of the lung and spleen homogenates on 7H10 agar plates containing 10% OADC, cycloheximide (10 ȝg/ml), and ampicillin (50 ȝg/ml). Plates were incubated for 4 weeks at 37°C before CFU were enumerated. CFU counts were log10 transformed before analysis as mean log10

CFU ± standard deviation (SD), and were compared by Student’s t-test using GraphPad Prism® version 7.0 software (GraphPad Software, Inc., La Jolla, CA, USA). P values less than 0.05 were considered as statistically significant.

Experiments were approved by the Swiss Cantonal Veterinary Authority (authorization no.

3082) and performed between June-August 2017.

215 ACKNOWLEDGMENTS

The research leading to these results was performed as part of the More Medicines for

Tuberculosis (MM4TB) project and received funding from the European Community’s Seventh

Framework Programme ([FP7/ 2007-2013]) under grant agreement no. 260872. We thank Aline

Reynaud and Florence Pojer from the Ppscf facility at EPFL for their technical support.

216 REFERENCES

1. Mikusova K, Huang H, Yagi T, Holsters M, Vereecke D, D'Haeze W, Scherman MS, Brennan PJ, McNeil MR, Crick DC. 2005. Decaprenylphosphoryl arabinofuranose, the donor of the D-arabinofuranosyl residues of mycobacterial arabinan, is formed via a two- step epimerization of decaprenylphosphoryl ribose. Journal of bacteriology 187:8020-5. 2. Brecik M, Centarova I, Mukherjee R, Kolly GS, Huszar S, Bobovska A, Kilacskova E, Mokosova V, Svetlikova Z, Sarkan M, Neres J, Kordulakova J, Cole ST, Mikusova K. 2015. DprE1 Is a Vulnerable Tuberculosis Drug Target Due to Its Cell Wall Localization. ACS Chem Biol 10:1631-6. 3. Piton J, Foo CS, Cole ST. 2016. Structural studies of Mycobacterium tuberculosis DprE1 interacting with its inhibitors. Drug Discov Today doi:10.1016/j.drudis.2016.09.014. 4. Makarov V, Manina G, Mikusova K, Mollmann U, Ryabova O, Saint-Joanis B, Dhar N, Pasca MR, Buroni S, Lucarelli AP, Milano A, De Rossi E, Belanova M, Bobovska A, Dianiskova P, Kordulakova J, Sala C, Fullam E, Schneider P, McKinney JD, Brodin P, Christophe T, Waddell S, Butcher P, Albrethsen J, Rosenkrands I, Brosch R, Nandi V, Bharath S, Gaonkar S, Shandil RK, Balasubramanian V, Balganesh T, Tyagi S, Grosset J, Riccardi G, Cole ST. 2009. Benzothiazinones kill Mycobacterium tuberculosis by blocking arabinan synthesis. Science 324:801-4. 5. Makarov V, Lechartier B, Zhang M, Neres J, van der Sar AM, Raadsen SA, Hartkoorn RC, Ryabova OB, Vocat A, Decosterd LA, Widmer N, Buclin T, Bitter W, Andries K, Pojer F, Dyson PJ, Cole ST. 2014. Towards a new combination therapy for tuberculosis with next generation benzothiazinones. EMBO Mol Med 6:372-83. 6. Batt SM, Jabeen T, Bhowruth V, Quill L, Lund PA, Eggeling L, Alderwick LJ, Futterer K, Besra GS. 2012. Structural basis of inhibition of Mycobacterium tuberculosis DprE1 by benzothiazinone inhibitors. Proc Natl Acad Sci U S A 109:11354-9. 7. Trefzer C, Rengifo-Gonzalez M, Hinner MJ, Schneider P, Makarov V, Cole ST, Johnsson K. 2010. Benzothiazinones: prodrugs that covalently modify the decaprenylphosphoryl- beta-D-ribose 2'-epimerase DprE1 of Mycobacterium tuberculosis. J Am Chem Soc 132:13663-5. 8. Neres J, Pojer F, Molteni E, Chiarelli LR, Dhar N, Boy-Rottger S, Buroni S, Fullam E, Degiacomi G, Lucarelli AP, Read RJ, Zanoni G, Edmondson DE, De Rossi E, Pasca MR,

217 McKinney JD, Dyson PJ, Riccardi G, Mattevi A, Cole ST, Binda C. 2012. Structural basis for benzothiazinone-mediated killing of Mycobacterium tuberculosis. Science translational medicine 4:150ra121. 9. Christophe T, Jackson M, Jeon HK, Fenistein D, Contreras-Dominguez M, Kim J, Genovesio A, Carralot JP, Ewann F, Kim EH, Lee SY, Kang S, Seo MJ, Park EJ, Skovierova H, Pham H, Riccardi G, Nam JY, Marsollier L, Kempf M, Joly-Guillou ML, Oh T, Shin WK, No Z, Nehrbass U, Brosch R, Cole ST, Brodin P. 2009. High content screening identifies decaprenyl-phosphoribose 2' epimerase as a target for intracellular antimycobacterial inhibitors. PLoS Pathog 5:e1000645. 10. Magnet S, Hartkoorn RC, Szekely R, Pato J, Triccas JA, Schneider P, Szantai-Kis C, Orfi L, Chambon M, Banfi D, Bueno M, Turcatti G, Keri G, Cole ST. 2010. Leads for antitubercular compounds from kinase inhibitor library screens. Tuberculosis (Edinb) 90:354-60. 11. Stanley SA, Grant SS, Kawate T, Iwase N, Shimizu M, Wivagg C, Silvis M, Kazyanskaya E, Aquadro J, Golas A, Fitzgerald M, Dai H, Zhang L, Hung DT. 2012. Identification of novel inhibitors of M. tuberculosis growth using whole cell based high-throughput screening. ACS Chem Biol 7:1377-84. 12. Landge S, Mullick AB, Nagalapur K, Neres J, Subbulakshmi V, Murugan K, Ghosh A, Sadler C, Fellows MD, Humnabadkar V, Mahadevaswamy J, Vachaspati P, Sharma S, Kaur P, Mallya M, Rudrapatna S, Awasthy D, Sambandamurthy VK, Pojer F, Cole ST, Balganesh TS, Ugarkar BG, Balasubramanian V, Bandodkar BS, Panda M, Ramachandran V. 2015. Discovery of benzothiazoles as antimycobacterial agents: Synthesis, structure- activity relationships and binding studies with Mycobacterium tuberculosis decaprenylphosphoryl-beta-D-ribose 2'-oxidase. Bioorg Med Chem 23:7694-710. 13. Landge S, Ramachandran V, Kumar A, Neres J, Murugan K, Sadler C, Fellows MD, Humnabadkar V, Vachaspati P, Raichurkar A, Sharma S, Ravishankar S, Guptha S, Sambandamurthy VK, Balganesh TS, Ugarkar BG, Balasubramanian V, Bandodkar BS, Panda M. 2016. Nitroarenes as Antitubercular Agents: Stereoelectronic Modulation to Mitigate Mutagenicity. ChemMedChem 11:331-9.

218 14. Foo CS, Lechartier B, Kolly GS, Boy-Rottger S, Neres J, Rybniker J, Lupien A, Sala C, Piton J, Cole ST. 2016. Characterization of DprE1-Mediated Benzothiazinone Resistance in Mycobacterium tuberculosis. Antimicrob Agents Chemother 60:6451-6459. 15. Wang F, Sambandan D, Halder R, Wang J, Batt SM, Weinrick B, Ahmad I, Yang P, Zhang Y, Kim J, Hassani M, Huszar S, Trefzer C, Ma Z, Kaneko T, Mdluli KE, Franzblau S, Chatterjee AK, Johnsson K, Mikusova K, Besra GS, Futterer K, Robbins SH, Barnes SW, Walker JR, Jacobs WR, Jr., Schultz PG. 2013. Identification of a small molecule with activity against drug-resistant and persistent tuberculosis. Proc Natl Acad Sci U S A 110:E2510-7. 16. Shirude PS, Shandil R, Sadler C, Naik M, Hosagrahara V, Hameed S, Shinde V, Bathula C, Humnabadkar V, Kumar N, Reddy J, Panduga V, Sharma S, Ambady A, Hegde N, Whiteaker J, McLaughlin RE, Gardner H, Madhavapeddi P, Ramachandran V, Kaur P, Narayan A, Guptha S, Awasthy D, Narayan C, Mahadevaswamy J, Vishwas KG, Ahuja V, Srivastava A, Prabhakar KR, Bharath S, Kale R, Ramaiah M, Choudhury NR, Sambandamurthy VK, Solapure S, Iyer PS, Narayanan S, Chatterji M. 2013. Azaindoles: noncovalent DprE1 inhibitors from scaffold morphing efforts, kill Mycobacterium tuberculosis and are efficacious in vivo. J Med Chem 56:9701-8. 17. Chatterji M, Shandil R, Manjunatha MR, Solapure S, Ramachandran V, Kumar N, Saralaya R, Panduga V, Reddy J, Prabhakar KR, Sharma S, Sadler C, Cooper CB, Mdluli K, Iyer PS, Narayanan S, Shirude PS. 2014. 1,4-azaindole, a potential drug candidate for treatment of tuberculosis. Antimicrob Agents Chemother 58:5325-31. 18. Neres J, Hartkoorn RC, Chiarelli LR, Gadupudi R, Pasca MR, Mori G, Venturelli A, Savina S, Makarov V, Kolly GS, Molteni E, Binda C, Dhar N, Ferrari S, Brodin P, Delorme V, Landry V, de Jesus Lopes Ribeiro AL, Farina D, Saxena P, Pojer F, Carta A, Luciani R, Porta A, Zanoni G, De Rossi E, Costi MP, Riccardi G, Cole ST. 2015. 2- Carboxyquinoxalines kill mycobacterium tuberculosis through noncovalent inhibition of DprE1. ACS Chem Biol 10:705-14. 19. Naik M, Humnabadkar V, Tantry SJ, Panda M, Narayan A, Guptha S, Panduga V, Manjrekar P, Jena LK, Koushik K, Shanbhag G, Jatheendranath S, Manjunatha MR, Gorai G, Bathula C, Rudrapatna S, Achar V, Sharma S, Ambady A, Hegde N, Mahadevaswamy J, Kaur P, Sambandamurthy VK, Awasthy D, Narayan C, Ravishankar S, Madhavapeddi

219 P, Reddy J, Prabhakar K, Saralaya R, Chatterji M, Whiteaker J, McLaughlin B, Chiarelli LR, Riccardi G, Pasca MR, Binda C, Neres J, Dhar N, Signorino-Gelo F, McKinney JD, Ramachandran V, Shandil R, Tommasi R, Iyer PS, Narayanan S, Hosagrahara V, Kavanagh S, Dinesh N, Ghorpade SR. 2014. 4-aminoquinolone piperidine amides: noncovalent inhibitors of DprE1 with long residence time and potent antimycobacterial activity. J Med Chem 57:5419-34. 20. Makarov V, Neres J, Hartkoorn RC, Ryabova OB, Kazakova E, Sarkan M, Huszar S, Piton J, Kolly GS, Vocat A, Conroy TM, Mikusova K, Cole ST. 2015. The 8-Pyrrole- Benzothiazinones Are Noncovalent Inhibitors of DprE1 from Mycobacterium tuberculosis. Antimicrob Agents Chemother 59:4446-52. 21. Shaikh MH, Subhedar DD, Arkile M, Khedkar VM, Jadhav N, Sarkar D, Shingate BB. 2016. Synthesis and bioactivity of novel triazole incorporated benzothiazinone derivatives as antitubercular and antioxidant agent. Bioorg Med Chem Lett 26:561-9. 22. Tiwari R, Miller PA, Chiarelli LR, Mori G, Sarkan M, Centarova I, Cho S, Mikusova K, Franzblau SG, Oliver AG, Miller MJ. 2016. Design, Syntheses, and Anti-TB Activity of 1,3-Benzothiazinone Azide and Click Chemistry Products Inspired by BTZ043. ACS Med Chem Lett 7:266-70. 23. Panda M, Ramachandran S, Ramachandran V, Shirude PS, Humnabadkar V, Nagalapur K, Sharma S, Kaur P, Guptha S, Narayan A, Mahadevaswamy J, Ambady A, Hegde N, Rudrapatna SS, Hosagrahara VP, Sambandamurthy VK, Raichurkar A. 2014. Discovery of pyrazolopyridones as a novel class of noncovalent DprE1 inhibitor with potent antimycobacterial activity. J Med Chem 57:4761-71. 24. Chikhale R, Menghani S, Babu R, Bansode R, Bhargavi G, Karodia N, Rajasekharan MV, Paradkar A, Khedekar P. 2015. Development of selective DprE1 inhibitors: Design, synthesis, crystal structure and antitubercular activity of benzothiazolylpyrimidine-5- carboxamides. Eur J Med Chem 96:30-46. 25. Amidon GL, Lennernas H, Shah VP, Crison JR. 1995. A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability. Pharm Res 12:413-20.

220 26. Daina A, Michielin O, Zoete V. 2017. SwissADME: a free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci Rep 7:42717. 27. Guidelines W. 2014. Guidelines on Post-Exposure Prophylaxis for HIV and the Use of Co- Trimoxazole Prophylaxis for HIV-Related Infections Among Adults, Adolescents and Children: Recommendations for a Public Health Approach: December 2014 supplement to the 2013 consolidated guidelines on the use of antiretroviral drugs for treating and preventing HIV infection, Geneva. 28. Bergstrom CA, Wassvik CM, Johansson K, Hubatsch I. 2007. Poorly soluble marketed drugs display solvation limited solubility. J Med Chem 50:5858-62. 29. Kabsch W. 2010. Integration, scaling, space-group assignment and post-refinement. Acta Crystallogr D Biol Crystallogr 66:133-44. 30. Adams PD, Afonine PV, Bunkoczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW, Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ, Richardson DC, Richardson JS, Terwilliger TC, Zwart PH. 2010. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr D Biol Crystallogr 66:213-21. 31. McCoy AJ, Grosse-Kunstleve RW, Adams PD, Winn MD, Storoni LC, Read RJ. 2007. Phaser crystallographic software. J Appl Crystallogr 40:658-674. 32. Emsley P, Cowtan K. 2004. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr 60:2126-32.

221 Table 1: Structure activity relationship of the different 2-sulphonylpiperazin 8-nitro 6- trifluoromethyl 1,3-benzothiazin-4-ones derivatives (sulfonyl PBTZ derivatives) in M. tuberculosis H37Rv

MIC99 MIC99 Molecular Molecular clogP H37Rv clogP H37Rv Compound structure Compound structure Name (μg/mL) Name (μg/mL)

O O S NO2 N S N S PBTZ169 4.31 0.0002 11326121 Cl 3.62 0.03 F N F F O

O O CH3 S NO N 2 N S N O CH Rifampicin 3.7 0.0008 11326056 3 1.79 0.06 F N F F O

O O O S CH 3 S NO2 N O NO2 N F F S N S N 11626093 2.69 0.001 11326120 O F 4.56 0.1 F N F N F F F O F O

O O O S S NO N NO N 2 2 O S N F S N 11626091 2 0.003 11326123 F 3.98 0.1 F N F N F F F F O F O CH O O 3 CH3 S O O O NO N 2 O S NO N S N 2 S 11626092 1.82 0.003 11326128 S N 2.59 0.1 F N F N F F F O F O

O O O O S S S NO2 N NO N 2 N S N S N CH N 3 CH 11326059 2.6 0.004 11326058 3 2.35 0.4 F N F N F F F O F O

O O CH S O O 3 S NO2 N NO2 N N S N S N N CH 11326127 3.28 0.006 11326061 3 1.81 0.4 F N F N F F F O F O

O O O S NO2 N S S N CH3 NO2 N O CH3 11626094 S N 3.1 0.01 11326125 5.23 0.5 F N CH3 F F N F O F F O O O CH3 O S S NO2 N NO N O 2 Cl S N S N S 11626095 1.36 0.02 11326122 Cl 4.51 0.8 F N F N F F F O F O

CH O O 3 S NO2 N O 11326126 S N N 1.33 1.6 F N F F O

O O S NO2 N S N CH 11326119 3 0.15 50 F N O F F O

O O S NO2 N 11326124 S N CH 5.06 800 3

F N F F O

222 Table 2: In vitro activity of selected sPBTZ derivatives against M. tuberculosis H37Rv and its BTZ-resistant mutant (NTB1) and cytotoxicity for HepG2 cells

H37Rv NTB1 Selective Compound TD50 MIC99 MIC99 index Name (μg/mL) (μg/mL) (μg/mL) 11326059 0.004 >100 100 25000 11326127 0.006 >100 >100 >16666 11626091 0.003 14.7 8.10 2700 11626092 0.003 25.1 8.90 2967 11626093 0.001 >100 >100 >100000 11626094 0.01 >100 100.00 10000 11626095 0.02 75 7.80 390 PBTZ169 0.0002 >100 44.00 220000 Rifampicin 0.0008 0.0008 100.00 125000

223 Table 3: Water solubility of selected sPBTZ derivatives measured by shake flask method compared to solubility calculated on SwissAdme ((26))

Compound Solubility measured Solubility calculated Name in water (μg/mL) in water (μg/ml) clogP

11326059 <0.01 0.12 2.6

11326127 <0.01 0.26 3.28 11626091 8.6 ± 3.8 4.69 1.91 11626092 10.7 ± 4.6 7.2 2 11626093 0.105 ± 0.05 0.93 2.69 11626094 0.088 ± 0.04 0.18 3.1 11626095 10.6 ± 1.8 16.9 1.36 PBTZ169 31.1 ± 6.4 0.0795 4.31

224 Table 4: Metabolic stability of sPBTZ derivatives measured in mouse and human microsomes

Clint mouse Clint human microsome Compound Name microsome References (ȝl/min/mg protein) (ȝl/min/mg protein)

11326059 18.6 16.8 This study

11326127 19 21.5 This study

11626091 2.7 9.7 This study

11626092 3 1.4 This study

11626093 42 6.4 This study

11626094 24 277 This study

11626095 0.9 1.0 This study

Carbamazepine 0.6 0.4 This study

Nifedipine 105 121.1 This study

PBTZ169 36.7 28 (5)

BTZ043 16.5 4.20 (5)

225

Figure 1: Structural comparison between the crystal structure of DprE1 in complex with sulfonyl derivative sPBTZ169 (11326127) and the structure in complex with MCZ (PDB code : 4NCR (5)). DprE1 is represented in green in the cartoon. sPBTZ169 (in red) sits in the trifluoromethyl hydrophobic pocket (in yellow) and binds covalently to C387. It is maintained by FAD (orange) and some lateral chains represented in sticks. Y60, a key residue in the binding of 2-Carboxyquinoxalines, is represented in pink.

226

Figure 2: Activity of the sPBTZ derivative 11626091 (50 mg/kg ) compared to in vivo activity of INH (10 mg/kg) or MCZ (25 mg/kg) in the mouse model of chronic TB. Black columns correspond to the bacterial burden in lungs and red columns correspond to the bacterial burden in the spleens, at day 0 (D0) when treatment was initiated, or day 28 (NT) when treatment ended. Bars represent the mean ± s.d. of CFUs from 5 mice per group. NT: untreated control. Limit of detection: 20 CFU/organ. The significance of difference was calculated using the Student t-test. *P<0.05. **P<0.005. ***P<0.0001 compared to NT.

227 SUPPLEMENTAL MATERIAL

Synthetic route for 2-[4-(R-sulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)- 4H-1,3-benzothiazin-4-one derivatives

O R S NO2 NH NO2 NO2 N O S SMe S N a) b) S N N N F C N F C 3 3 F3C O O O 1 2 3a-s 3R 3R a k Me Me N Me O b l Et N Me N Et c m

S d n n-butyl N Me N Me e o

S Cl f p Cl

S Cl g q

CF3 N Me O h r

OCF3 MeOOC S i O s

Me N j Me Et Me

228 General Procedure for the synthesis of 2-[4-(R-sulfonyl)piperazin-1-yl]-8-nitro-6- (trifluoromethyl)-4H-1,3-benzothiazin-4-one derivatives. Step a). At 20 oC, finely ground solid 2-(methylthio)-8-nitro-6-(trifluoromethyl)-4H- 1,3-benzothiazin-4-one (25 mmol) (Makarov, EMBO Mol Med, 2014) was added in one portion under intensive mixing to solution of piperazine hexahydrate (125 mmol) in 15 ml of ethanol. After 5 min reaction mixture was diluted by water, yellow solid was filtered off and recrystallised from ethyl acetate twice. The yield of 8-nitro-2- piperazin-1-yl-6-(trifluoromethyl)-4H-1,3-benzothiazin-4-one is 74%, mp 151-5 oC. + 1 Mass (EI), m/z (Irelat.(%)): 359.3126 [M] (45). C11H13F3N4O3S. H NMR (DMSO-d6):

3.16 (4H, br s, N(CH2)2), 3.98 (4H, br s, N(CH2)2), 8.73 (1H, s, CH), 8.87 (1H, s, CH) ppm. Step b). Solution of 8-nitro-2-piperazin-1-yl-6-(trifluoromethyl)-4H-1,3-benzothiazin- 4-one (10 mmol) in 5 mL of pyridine was treated by correspomding sulfonyl chloride (10 mmol) at 20 oC. After 1 hour the reaction mixture was diluted by water and acidified by concentrated hydrochloride solution in water till pH 1. Formed solid of 2-[4-(R- sulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3-benzothiazin-4-one was filtered off and washed by water. 2-[4-(Methylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3-benzothiazin-

4-one 3a (11626095) Yield 61%. Mp. 186- ɋ (W2+ 0DVV (, m/z (Irelat.(%)): + 1 437.4042 [M] (49). C14H13F3N4O5S2. H NMR (DMSO-d6): 2.93 (3H, s, CH3), 3.24

(4H, br s, N(CH2)2), 4.05 (4H, br s, N(CH2)2), 8.76 (1H, s, CH), 8.88 (1H, s, CH) ppm. 2-[4-(Ethylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3-benzothiazin-

4-one 3b (11626092) Yield 74%. Mp. 204-ɋ (W2+ 0DVV (, m/z (Irelat.(%)): + 1 451.4208 [M] (86). C15H15F3N4O5S2. H NMR (DMSO-d6): 1.34 (3H, t, J = 7.4 Hz,

CH3), 2.75 (2H, q, J = 7.4 Hz, CH2), 3.26 (4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 8.77 (1H, s, CH), 8.85 (1H, s, CH) ppm. 2-[4-(Cyclopropylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3c (11626091) Yield 57%. Mp. 201-ɋ (W2+ 0DVV (, m/z + 1 (Irelat.(%)): 463.4415 [M] (52). C16H15F3N4O5S2. H NMR (DMSO-d6): 1.48 (4H, m,

CH2-CH2), 3.26 (4H, br s, N(CH2)2), 3.81 (1H, m, CH), 4.09 (4H, br s, N(CH2)2), 8.73 (1H, s, CH), 8.86 (1H, s, CH) ppm. 2-[4-(n-butylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3d (11626093) Yield 64%. Mp. 215-ɋ +2O/EtOH). Mass (EI), + 1 m/z (Irelat.(%)): 479.4839 [M] (89). C17H19F3N4O5S2. H NMR (DMSO-d6): 0.85 (3H,

229 t, J = 7.2 Hz, CH3), 1.23 (2H, m, CH2), 1.66 (2H, m, CH2), 2.64 (2H, t, J = 7.2 Hz,

CH2), 3.21 (4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 8.75 (1H, s, CH), 8.84 (1H, s, CH) ppm. 2-[4-(Cyclohexylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3e (11326127) Yield 47%. Mp. 283-ɋ (W2+ 0DVV (, m/z + 1 (Irelat.(%)): 463.4415 [M] (72). C19H21F3N4O5S2. H NMR (DMSO-d6): 1.52 (6H, m,

(CH2)3), 2.42 (4H, m, 2CH2), 2.64 (2H, t, J = 7.2 Hz, CH2), 3.21 (4H, br s, N(CH2)2),

4.06 (4H, br s, N(CH2)2), 8.75 (1H, s, CH), 8.84 (1H, s, CH) ppm. 2-{4-[(4-butylphenyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3f (11326124) Yield 37%. Mp. 226-ɋ (W2+ 0DVV (, m/z + 1 (Irelat.(%)): 555.5799 [M] (69). C23H23F3N4O5S2. H NMR (DMSO-d6): 0.88 (3H, t, J

= 7.2 Hz, CH3), 1.23 (2H, m, CH2), 1.66 (2H, m, CH2), 2.66 (2H, m, CH2), 3.21 (4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 7.03 (2H, d, J = 8.7 Hz, 2CH), 7.66 (2H, d, J = 8.7 Hz, 2CH), 8.77 (1H, s, CH), 8.86 (1H, s, CH) ppm. 2-{4-[(4-trifluoromethylphenyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)- 4H-1,3-benzothiazin-4-one 3g (11326123) Yield 46%. Mp. 222- ɋ (W2+ 0DVV + 1 (EI), m/z (Irelat.(%)): 567.4715 [M] (50). C20H14F6N4O5S2. H NMR (DMSO-d6): 3.22

(4H, br s, N(CH2)2), 4.05 (4H, br s, N(CH2)2), 7.82 (2H, d, J = 6.3 Hz, 2CH), 8.04 (2H, d, J = 6.3 Hz, 2CH), 8.72 (1H, s, CH), 8.88 (1H, s, CH) ppm. 2-{4-[(4-trifluoromethoxyphenyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)- 4H-1,3-benzothiazin-4-one 3h (11326120) Yield 52%. Mp. 207- ɋ (W2+ 0DVV + 1 (EI), m/z (Irelat.(%)): 583.4709 [M] (36). C20H14F6N4O7S2. H NMR (DMSO-d6): 3.20

(4H, br s, N(CH2)2), 4.01 (4H, br s, N(CH2)2), 7.62 (2H, d, J = 6.3 Hz, 2CH), 7.90 (2H, d, J = 6.3 Hz, 2CH), 8.71 (1H, s, CH), 8.83 (1H, s, CH) ppm. 2-{4-[(4-acetylphenyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3i (11326119) Yield 64%. Mp. 280-ɋ &+3CN). Mass (EI), m/z + 1 (Irelat.(%)): 541.5102 [M] (59). C21H17F3N4O6S2. H NMR (DMSO-d6): 2.48 (3H, s,

CH3), 3.26 (4H, br s, N(CH2)2), 4.03 (4H, br s, N(CH2)2), 7.97 (4H, m, 4CH), 8.71 (1H, s, CH), 8.83 (1H, s, CH) ppm. 2-(4-{[4-(1,1-dimethylpropyl)phenyl]sulfonyl}piperazin-1-yl)-8-nitro-6- (trifluoromethyl)-4H-1,3-benzothiazin-4-one 3j (11326125) Yield 37%. Mp. 231-ɋ + 1 (EtOH). Mass (EI), m/z (Irelat.(%)): 569.6065 [M] (33). C24H25F3N4O6S2. H NMR

(DMSO-d6): 0.76 (3H, t, J = 7.2 Hz, CH3), 1.31 (6H, s, 2CH3), 1.69 (2H, m, CH2), 3.23

230 (4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 7.33 (2H, d, J = 8.7 Hz, 2CH), 7.47 (2H, d, J = 8.7 Hz, 2CH), 8.73 (1H, s, CH), 8.87 (1H, s, CH) ppm. 2-{4-[(3,5-dimethylisoxazol-4-yl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)- 4H-1,3-benzothiazin-4-one 3k (11326056) Yield 79%. Mp. 265- ɋ (W2+'0) + 1 Mass (EI), m/z (Irelat.(%)): 518.4813 [M] (31). C18H16F3N5O6S2. H NMR (DMSO- d6): 2.31 (3H, s, CH3), 2.64 (3H, s, CH3), 3.26 (4H, br s, N(CH2)2), 4.05 (4H, br s,

N(CH2)2), 8.77 (1H, s, CH), 8.85 (1H, s, CH) ppm. 2-{4-[(1-ethyl-3-methyl-1H-pyrazol-4-yl)sulfonyl]piperazin-1-yl}-8-nitro-6- (trifluoromethyl)-4H-1,3-benzothiazin-4-one 3l (11326058) Yield 73%. Mp. 218-ɋ + 1 (EtOH). Mass (EI), m/z (Irelat.(%)): 531.5237 [M] (44). C19H19F3N6O5S2. H NMR

(DMSO-d6): 1.31 (3H, t, J = 5.9 Hz, CH3), 2.28 (3H, s, CH3), 3.18 (4H, br s, N(CH2)2),

4.06 (4H, br s, N(CH2)2), 4.09 (2H, q, J = 6.6 Hz, CH3), 8.27 (1H, s, CH), 8.78 (1H, s, CH), 8.86 (1H, s, CH) ppm. 8-nitro-2-[4-(2-thienylsulfonyl)piperazin-1-yl]-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3m (11326059) Yield 80%. Mp. 210-ɋ (W2+ 0DVV (, m/z + 1 (Irelat.(%)): 505.5042 [M] (67). C17H13F3N4O5S3. H NMR (DMSO-d6): 3.20 (4H, br s, N(CH2)2), 4.05 (4H, br s, N(CH2)2), 7.25 (1H, br s, CH), 7.68 (1H, br s, CH), 8.03 (1H, br s, CH), 8.73 (1H, s, CH), 8.83 (1H, s, CH) ppm. 2-{4-[(1,3-dimethyl-1H-pyrazol-4-yl)sulfonyl]piperazin-1-yl}-8-nitro-6- (trifluoromethyl)-4H-1,3-benzothiazin-4-one 3n (11326061) Yield 64%. Mp. 200-ɋ + 1 (EtOH). Mass (EI), m/z (Irelat.(%)): 517.4966 [M] (74). C18H17F3N6O5S2. H NMR

(DMSO-d6): 2.41 (3H, s, CH3), 3.17 (4H, br s, N(CH2)2), 3.72 (3H, s, NCH3), 4.03 (4H, br s, N(CH2)2), 7.63 (s, 1H, CH), 8.75 (1H, s, CH), 8.86 (1H, s, CH)ppm. 2-{4-[(5-chloro-2-thienyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothia-zin-4-one 3o (11326121) Yield 39%. Mp. 210-ɋ (W2+'0) 0DVV (, + 1 m/z (Irelat.(%)): 539.9492 [M] (53). C17H12F3ClN4O5S3. H NMR (DMSO-d6): 3.24

(4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 7.39 (1H, d, J = 5.1 Hz, CH), 7.58 (1H, d, J = 5.1 Hz, CH),, 8.78 (1H, s, CH), 8.85 (1H, s, CH) ppm. 2-{4-[(4,5-dichloro-2-thienyl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)-4H- 1,3-benzothia-zin-4-one 3p (11326122) Yield 46%. Mp. 265-ɋ (W2+'0) 0DVV + 1 (EI), m/z (Irelat.(%)): 574.3942 [M] (58). C17H11F3Cl2N4O5S3. H NMR (DMSO-d6):

3.26 (4H, br s, N(CH2)2), 4.07 (4H, br s, N(CH2)2), 7.83 (1H, s, CH), 8.77 (1H, s, CH), 8.86 (1H, s, CH) ppm.

231 2-{4-[(3-methylisoxazol-4-yl)sulfonyl]piperazin-1-yl}-8-nitro-6-(trifluoromethyl)-4H- 1,3-benzothiazin-4-one 3q (11326126) Yield 86%. Mp. 281-ɋ (W2+'0) 0DVV + 1 (EI), m/z (Irelat.(%)): 504.4542 [M] (61). C17H14F3N5O6S2. H NMR (DMSO-d6): 2.33

(3H, s, CH3), 3.26 (4H, br s, N(CH2)2), 4.04 (4H, br s, N(CH2)2), 7.69 (1H, s, CH), 8.75 (1H, s, CH), 8.86 (1H, s, CH) ppm. Methyl 3-({4-[8-nitro-4-oxo-6-(trifluoromethyl)-4H-1,3-benzothiazin-2-yl]piperazin- 1-yl}sulfo-nyl)thiophene-2-carboxylate 3r (11326128) Yield 74%. Mp. 127- ɋ + 1 (EtOH/DMF). Mass (EI), m/z (Irelat.(%)): 563.5412 [M] (68). C19H15F3N4O7S3. H

NMR (DMSO-d6): 3.20 (4H, br s, N(CH2)2), 3.83 (3H, s, OCH3), 4.07 (4H, br s,

N(CH2)2), 6.81 (1H, d, J = 5.3 Hz, CH), 7.67 (1H, d, J = 5.3 Hz, CH), 8.70 (1H, s, CH), 8.84 (1H, s, CH) ppm. 2-[4-(isoquinolin-5-ylsulfonyl)piperazin-1-yl]-8-nitro-6-(trifluoromethyl)-4H-1,3- benzothiazin-4-one 3s (11626094) Yield 85%. Mp. 243-ɋ (W2+'0) 0DVV (, + 1 m/z (Irelat.(%)): 550.5265 [M] (36). C22H16F3N5O5S2. H NMR (DMSO-d6): 3.24 (4H, br s, N(CH2)2), 4.06 (4H, br s, N(CH2)2), 7.80 (1H, t, J = 7.9 Hz, CH), 7.99 (1H, d, J = 5.6 Hz, CH), 8.38 (2H, m, 2CH), 8.73 (1H, s, CH), 8.85 (1H, s, CH), 8.93 (1H, d, J = 5.6 Hz, CH), 9.57 (1H, s, CH) ppm.

232 Table S1 : IC50 values measured for Ty38c, PBTZ169 and sPBTZ169 (11326127) against wild-type and BTZ-resistant M. tuberculosis C387S DprE1

IC50 (μM) compound structure References WT C387S

PBTZ169 0.3 3.6 This study

Ty38c 0.2 0.2 (1)

O O S NO2 N S N

sPBTZ169 F N 1.1 12 This study F F O

233 Table S2. Data collection and refinement statistics

DprE1-sPBTZ169 Beamline SLS X06DA (PXIII) Wavelength (Å) 1 Resolution range (Å) 49.15- 2.40 (2.48-2.40)

Space group P 1 21 1 78.164 84.024 81.666 Unit cell (Å, °) 90 103.983 90 Total reflections 79124 (7781) Unique reflections 39831 (3914) Multiplicity 2.0 (2.0) Completeness (%) 98.81 (98.19) Mean I/sigma(I) 16.42 (2.56) R-merge 0.03362 (0.3709) R-meas 0.04754 (0.5245) CC1/2 0.999 (0.85) CC* 1 (0.959) R-work 0.1978 R-free 0.2248 Number of non-hydrogen 6512 atoms Ramachandran favored (%) 97.58 Ramachandran allowed (%) 2.42 Ramachandran outliers (%) 0.00 Rotamer outliers (%) 1.92 Clashscore 4.47 Average B-factor 60.34

Statistics for the highest-resolution shell are shown in parentheses.

234 Fig S1: Antitubercular activity versus hydrophobicity. Distribution and activity of 2-sulphonylpiperazin 8-nitro 6-trifluoromethyl 1,3-benzothiazin-4-ones derivatives as a function of clogP. Surrounded in red, the 7 chosen compounds with a good activity/logP profile. Lozenges indicate other compounds not discussed in the text.

References

1. Foo CS, Lechartier B, Kolly GS, Boy-Rottger S, Neres J, Rybniker J, Lupien A, Sala C, Piton J, Cole ST. 2016. Characterization of DprE1-Mediated Benzothiazinone Resistance in Mycobacterium tuberculosis. Antimicrob Agents Chemother 60:6451-6459.

235

Appendix Chapter A3: An optimized background regimen for treatment of active tuberculosis with the next-generation benzothiazinone Macozinone (PBTZ169)

Andréanne Lupien1, Anthony Vocat1, Caroline Shi-Yan Foo, Emilyne Blattes2, Jean-Yves Gillon2*, Vadim Makarov3, Stewart T. Cole1,†

1Global Health Institute, Ecole Polytechnique Fédérale de Lausanne, CH-1015 Lausanne, Switzerland 2Innovative Medicines for Tuberculosis (iM4TB), Innovation Park, Building I, CH-1015 Lausanne, Switzerland 3Dept. of Stresses of Microorganisms, A.N. Bach Institute of Biochemistry, Moscow, Russian Federation *Present address: Drugs for Neglected Diseases initiative, 15 Chemin Louis-Dunant,1202 Geneva, Switzerland

Antimicrobial Agents and Chemotherapy, 2018, doi: 10.1128/AAC.00840-18

Contributions: execution of viability assays, manuscript preparation

237

ABSTRACT

The efficacy of the standardized four-drug regimen (comprising isoniazid, rifampicin, pyrazinamide and ethambutol) for the treatment of tuberculosis (TB) is menaced by the emergence of multidrug (MDR) and extensively drug-resistant (XDR) strains of Mycobacterium tuberculosis.

Intensive efforts have been made to develop new antibiotics or to repurpose old drugs and several of these are currently being evaluated in clinical trials for their anti-tubercular activity. Among the new candidate drugs is macozinone (MCZ), the piperazine-containing benzothiazinone, PBTZ169, which is currently being evaluated in phase I/II clinical trials.

Here, we determined the in vitro and in vivo activity of MCZ in combination with a range of anti-TB drugs in order to design a new regimen against active TB. Two-drug combinations with

MCZ were tested against M. tuberculosis using checkerboard and CFU enumeration after drug exposure assays. MCZ was observed to have no interactions with all first- and second-line anti-TB drugs. At the MIC of each drug, MCZ with either bedaquiline (BDQ), clofazimine (CLO), delamanid (DMD) or sutezolid (STZ) reduced the bacterial burden by two logs compared to the drugs alone, indicating synergism. MCZ also displays synergism with clomiphene (CLM), a potential inhibitor of the undecaprenyl pyrophosphate synthase (UppS) in mycobacteria. For all the other drugs tested in combination with MCZ, no synergistic activity was observed. Neither antagonism nor increased cytotoxicity were found for most combinations, suggesting that MCZ could be added to different TB regimens without any significant adverse effects.

239 INTRODUCTION

With more than 9 million people infected worldwide, tuberculosis (TB) is the major cause of death due to a bacterial pathogen (1). As recommended by the WHO, the mainstay regimen for drug-susceptible TB relies on a multi-drug therapy which consists of isoniazid (INH), rifampicin

(RIF), pyrazinamide (PZA) and ethambutol (EMB) (1). The increasing number of multi-drug resistant TB (MDR-TB) and extensively resistant TB (XDR-TB) cases challenge the reliability of this regimen to treat active TB. MDR-TB strains are resistant to both isoniazid (INH) and rifampicin (RIF), the two most active drugs for drug-susceptible TB (DS-TB). MDR-TB treatment involves second-line injectable aminoglycosides (e.g. amikacin and capreomycin) and fluoroquinolones. In the case of XDR-TB, resistance to RIF, INH, aminoglycosides, and fluoroquinolones is observed. With more than 480 000 reported cases of MDR-TB and 9.5% of

MDR-TB cases worldwide classified as XDR-TB, drugs active against new Mycobacterium tuberculosis pharmacological targets are indeed in need (1).

An effective regimen for the treatment of TB consists of multiple drugs that inhibit several distinct essential cellular activities in the bacterium. The effectiveness of a regimen is further enhanced if synergism is present between the constituents. Recently, studying the combinatory effect of anti-TB agents has become of interest to the TB community in order to find more efficacious treatments. Several combinations of anti-TB drugs have demonstrated potentiation of the individual drug activity in vitro when combined. Interestingly, drugs such as cephalosporins, which have limited activity in M. tuberculosis due to the presence of the BlaC ȕ-lactamase, were shown to be synergistic with drugs of the rifamycin family, bedaquiline (BDQ) and delamanid

(DMD) (2, 3). Furthermore, the ȕ-lactamase inhibitor, clavulanate as a component of triple combinations including cephalosporins (and other beta-lactams) proved to be a key synergistic

240 partner due to its ability to rescue the activity of RIF against a RIF-resistant strain (2). A combination of amoxicillin/clavulanate with first-line anti-TB drugs has also been explored and showed potent synergism against MDR isolates (4). Repurposing drugs such as spectinomycin was also studied to test the concept that clinically available antibiotics with limited efficacy against M. tuberculosis might be used for TB treatment when co-administered with macrolides and azoles (5).

The combinatory effect of new drugs has also been assessed, whereby MmpL3 inhibitors act synergistically with RIF, BDQ, clofazimine (CLO)DQGȕ-lactams (6). Altogether, these studies demonstrate the potential of anti-TB drugs to act synergistically and the possibility to optimize treatment regimens against DS-TB, MDR-TB, and XDR-TB or even to develop a pan-TB regimen

(7).

A potential component of such regimens is the clinical candidate macozinone (MCZ) previously known as PBTZ169 (https://www.newtbdrugs.org/). MCZ is a bactericidal benzothiazinone that inhibits the essential flavoprotein DprE1 by forming a covalent bond with the active site Cys387 residue (8) thus preventing synthesis of decaprenyl phosphoryl ribose. Despite its lack of activity against non-replicating bacteria, MCZ is highly potent against M. tuberculosis in vitro (9), ex vivo and in vivo, where, alone, it significantly reduces the bacterial burden in the lungs and spleens of chronically infected mice (10). A new regimen including MCZ, BDQ, and

PZA was found to be more efficacious than the standard three-drug treatment in a murine model of chronic disease (10). MCZ was also shown to act synergistically with BDQ and CLO (10, 11).

BDQ is an inhibitor of the F1/F0 ATP synthase and CLO is thought to be reduced by the mycobacterial type 2 NADH:quinone oxidoreductase, NDH-2, leading to a CLO-mediated increase in NADH oxidation and production of reactive oxygen species (ROS) (12, 13). Other studies have

241 also reported that CLO may act on the lipids of the bacterial cell wall, as observed in macrophages infected with M. leprae (14). However, the mechanism of action of this drug is far from understood.

As part of a new regimen, further studies are needed to evaluate the potential activity of

MCZ when combined with other anti-TB agents. Although several new or repurposed drugs have been evaluated for their potential to be used with MCZ in combination against active TB, there has been no comprehensive work performed to optimize MCZ-containing regimens. In this study, we embarked on a thorough characterization and evaluation of the potential activity of MCZ in combination with other TB drugs using a combination of checkerboard and CFU enumeration after drug exposure assays.

242 RESULTS

Activity of MCZ when combined with first- and second-line anti-TB drugs against strain

H37Rv.

The MIC of all compounds used in this study against the H37Rv strain of M. tuberculosis is shown in Table 1 and was determined by a resazurin-based microtiter assay (REMA) (15). All drugs tested are active against the bacterium under these conditions except PZA. The activity of

MCZ was then determined in combination with the first-line drugs RIF, INH and EMB using checkerboard assay. At concentrations ranging from 0.5X MIC to 0.0625X MIC, the combination of MCZ with first-line drugs resulted in minimal fractional inhibitory index (FICIm) and maximal fractional inhibitory index (FICIM) values ranging from 0.75 to 2.0, respectively (Table 2). MCZ was also tested in combination with the second-line injectable agent amikacin (AMK), the fluoroquinolones, levofloxacin (LVX) and moxifloxacin (MFX), the bacteriostatic agents, D- cycloserine (DCS), ethionamide (ETO), and para-aminosalicylic acid (PAS). For all combinations,

FICIm and FICIM range from 1.125 to 3, respectively (Table 2). Taken together, these results show that MCZ has neither synergistic nor antagonistic interactions with the first- and second-line anti-

TB drugs.

Activity of MCZ with repurposed or new anti-TB drugs.

In cases of MDR-and XDR-TB disease, the use of new anti-TB drugs or repurposed drugs is required (16). Several repurposed drugs, including the anti-leprosy agent CLO, the macrolide clarithromycin (CLR), the carbapenem meropenem (MEM) and the oxazolidinone linezolid (LZD) are now in trials for the treatment of MDR-TB (17–20). In addition, new anti-TB compounds are available. These drugs include the diarylquinoline BDQ, the dihydro-nitroimidazole DMD, the

243 oxazolidinone STZ and the QcrB inhibitor lansoprazole sulfide LPZs. CLR, DMD, LPZs, LZD,

MEM, and STZ are not synergistic when individually combined with MCZ, with FICIm and FICIM ranging between 1.033 and 2.5 (Table 3). As previously reported, benzothiazinones have a synergistic effect when combined with CLO and BDQ (10, 11). In the checkerboard assay, synergism was observed at concentrations close to the MIC in vitro (0.25X MIC and above) suggesting that MCZ synergism with CLO and BDQ is dependent on the concentration of MCZ.

Interestingly, both BDQ and CLO were reported to act on M. tuberculosis by disrupting the proton motive force of the mycobacterial membrane (21). To assess the activity of MCZ with other uncouplers, we tested pairwise combinations of clomiphene (CLM) and tamoxifen (TMX), using the H37Rv-lux strain for TMX, due to a high fluorescence background at concentrations higher than the MIC when TMX susceptibility was determined by REMA (data not shown) (21). SQ109, although reported to act as an uncoupler, was not included in the study as its activity was previously shown to be indifferent in the presence of benzothiazinones (10, 21, 22). CLM and TMX alone are both active against H37Rv with a MIC of 24.8 ȝg/mL and 14.22 ȝg/mL respectively (Table 1).

Synergism was observed when MCZ was combined in vitro with CLM (Table 3) whereas no combinatory effect was found with TMX (Table 3).

CFU enumeration after drug exposure.

MCZ is a bactericidal drug that inhibits the DprE1 enzyme (10, 23). To assess the cidality of MCZ in two-drug combinations in vitro, CFU enumeration was performed on H37Rv in the presence of BDQ, CLO, MFX, DMD and STZ (Fig. 1B and Table S1). First, the activity of the drugs was assessed at 2X MIC, 1X MIC, 0.5X MIC and 0.25X MIC of each compound. The activity of INH and RIF was also determined at 1X MIC and 0.25X MIC, as controls (Fig. 1A). RIF and

244 INH decreased significantly the bacterial burden, by 2.2 log10 CFU/mL and below the detection limit, respectively (Fig. 1A), when compared to the non-treated control at D7. MCZ, BDQ, CLO,

DMD, and MFX showed dose-dependent bactericidal activity (Fig. 1B). No dose-dependency was observed when H37Rv was exposed to STZ (Fig. 1B). At the MIC of each drug, the combination of MCZ with BDQ, CLO, MOX, DMD, and STZ significantly decreased the number of CFU compared to each drug alone (Fig. 1C). In combination, MCZ significantly increased the bactericidal activity reducing the number of CFU below that seen before treatment (D0, Fig. 1C).

In the absence of MCZ, only the combination of CLO and BDQ had this effect (data not shown).

At 0.25X MIC of each drug, a significant decrease in the bacterial viability was only observed when MCZ was combined with CLO (Fig. 1C). Also, a decrease of 1.08 log10 CFU/mL was observed when MCZ was combined with BDQ at 0.25X MIC of each drug (Fig. 1C). This result suggests that despite the lack of synergism at 0.25X MIC, MCZ and BDQ can potentiate their bactericidal action in vitro. Interestingly, these results imply that at optimal concentrations of each drug, either combination of MCZ with BDQ, CLO, DEL, MFX or with STZ can increase the bactericidal activity in vitro compared to the drugs used alone.

Cytotoxicity of potent synergistic combinations containing MCZ.

Cytotoxicity of the synergistic combinations was determined against HepG2 cells. All drugs alone, except CLO, had TD50 and TD99 higher than 20 ȝg/mL (Table 4). The toxicity of two-drug combinations was similar to that of the more toxic drug in the combination except for

MCZ-DMD and DMD-STZ (Table 4). The TD50 and TD99 of the MCZ-DMD combination were

7.7-fold and 7.8-fold higher than MCZ and DMD alone, respectively. In the case of DMD-STZ,

245 the combination was 2.1-fold more toxic than STZ alone. Interestingly, the three-drug combination

MCZ-BDQ-DMD has similar toxicity to that of the MCZ-DMD combination.

In vivo activity of MCZ in combination with DMD and STZ.

Despite the lack of synergism in the checkerboard assay (Table 3), both MCZ-DMD and

MCZ-STZ combinations were able to decrease bacterial viability in vitro by more than 2 log10 CFU when added at a concentration equal to the MIC of each drug (Fig. 1C). Thus, the in vivo activity of these two drugs was evaluated in a TB mouse model (Fig. 2). At the concentrations tested, all drugs alone and in combinations significantly decreased the bacterial burden, by more than 2.18 log10 in the lungs and 1.71 log10 in the spleens, as compared to the D0 level, thus indicating a bactericidal effect. Although no combinatory effect was observed in these organs when MCZ,

DMD, and STZ were combined in two-drug combinations, the triple combination, MCZ-DMD-

STZ, was more active in the lungs of M. tuberculosis-infected mice than DMD-STZ (the most active of the two-drug combinations) with cidal activity of more than 0.69 log10 CFU (p-value =

0.014 (*). Moreover, MCZ-DMD-STZ was also more active than the mainstay regimen (RIF-INH-

PZA-RHZ) in this model, reducing the burden by 1.17 log10 CFU compared to RHZ (p-value =

0.004 (**). In the spleen, the most active two-drug combination was MCZ-DMD although no

VLJQLILFDQWGHFUHDVH 0.5log10 CFU) in the bacterial burden could be observed in the case of the triple combination of MCZ-DMD-STZ. These results suggest that in spite of the lack of synergism observed between MCZ, DMD, and STZ in this model, under the conditions tested, the MCZ-

DMD-STZ combination is efficacious in treating TB infection in the mouse, comparable to the mainstay regimen of RHZ.

246 DISCUSSION

In this study, we determined the combinatory effects of MCZ with not only first- and second-line TB drugs, but also with new and repurposed drugs with anti-tubercular activity.

Synergism is indeed of interest for the optimization of a multi-drug regimen, as the activity of the drugs would be potentiated by the presence of another compound. More importantly, antagonistic interactions between drugs should be avoided, so as not to impair the activity of the regimen. We demonstrated that MCZ has no antagonistic effects with the first-line anti-TB drugs RIF, INH and

EMB (Table 2) as well as the second-line drugs (Table 2). This is consistent with previously published results obtained in vitro with combinations containing BTZ043, the lead compound from the benzothiazinone chemical class (23). MCZ was previously shown to be synergistic with BDQ and CLO in vitro and in vivo (10, 11). In this study, we observed similar synergism between MCZ,

BDQ, and CLO in the checkerboard assay and enhanced bactericidal activity of MCZ in the CFU enumeration after drug exposure assay in vitro (Table 3 and Fig. 1).

Interestingly, MCZ is also synergistic with CLM in the in vitro checkerboard assay. CLM is a nonsteroidal, ovulatory stimulant that acts as a selective estrogen receptor modulator (24).

Little is known about the mechanism of action of CLM in M. tuberculosis. However, a previous study in Staphylococcus aureus showed that CLM targets the undecaprenyl diphosphate synthase

(UppS), an enzyme, which catalyzes the condensation of isopentenyl diphosphate with allylic pyrophosphates (25) and acts in the same pathway as DprE1 in actinobacteria. In Corynebacterium glutamicum, overexpression of UppS was able to rescue the wild-type strain and induce BTZ043 resistance by providing sufficient decaprenyl phosphate for cell wall biosynthesis. The 2.15 Å crystal structure (PDB: 5CQJ) of Escherichia coli UppS in the presence of CLM has already been determined (25). Despite the low identity between the M. tuberculosis and E. coli UppS sequences

247 (39.02%), 10 of the 16 amino acids reported to interact with CLM are conserved (M25, N28, H43,

G46, V50, A69, L93, L100, L107 and L139 according to E. coli numbering), suggesting that in M. tuberculosis CLM is also an inhibitor of UppS (Fig. S2). CLM may thus be used as a booster of

MCZ activity by further inhibiting the decaprenyl phosphate pathway of actinobacteria.

CLM, BDQ and CLO have all been reported to act as protonophore uncouplers which disrupt mycobacterial membrane potential (21). BDQ and CLO similarly demonstrated synergism with MCZ, as with CLM, in our study, although no synergism was observed with other uncouplers such as SQ109 and TMX (Table 3; (10)). SQ109 was reported to dissipate the transmembrane

SURWRQFRQFHQWUDWLRQJUDGLHQW ǻS+ PHPEUDQHSRWHQWLDO ǻȥ RUERWKFRPSRQHQWVRIWKHproton motive force simultaneously, mainly due to its inhibitory effect on MmpL3 and other MmpL-driven processes (22). Little is known about the mechanism of action of TMX against M. tuberculosis.It was previously hypothesized that TMX may cause a loss of transmembrane potential and ultimately cell death by a significant efflux of K+ and Na+ ions from bacterial cells (26) and this may also occur in M. tuberculosis. Interestingly, BDQ and CLO, which synergize with MCZ, act on components of the electron-transport chain (ETC) in the bacterial membrane. BDQ is an inhibitor of the F1/F0 ATP synthase and CLO is thought to be reduced by the mycobacterial type 2 NADH: quinone oxidoreductase, NDH-2 enzyme, leading to a CLO-mediated increase in NADH oxidation and reactive oxygen species (ROS) production (12, 13). Other studies have also reported that CLO may act on the lipids of the bacterial cell wall, as observed in macrophages infected with M. leprae

(14). On the other hand, there are compounds targeting the ETC such as LPZs, which are not synergistic when combined with MCZ in vitro (Table 3). LPZs targets the QcrB component of the cytochrome bc1 oxidase, loss of which is thought to be compensated for by the other terminal

248 oxidase in M. tuberculosis, cytochrome bd (27). Further studies will be needed to fully understand the mechanism underlying the two-drug synergism of MCZ with BDQ, CLO, and CLM.

As observed from the CFU enumeration after drug exposure assay at certain optimal concentrations, the bactericidal activity of MCZ is enhanced by BDQ, CLO, DMD, MFX, and STZ

(Fig. 1). In addition, BDQ and CLO potentiate their activity. Interestingly, all drugs tested were reported to have bactericidal activity against M. tuberculosis in vitro or in vivo (28–32).

Background regimens with MCZ and BDQ, CLO, DMD, MFX, or STZ would thus be promising alternatives to decrease the bacterial burden in vivo. In this study, MCZ in combination with DMD and STZ was assessed in a murine infection model. Despite the lack of synergism between the drugs in vivo, this regimen was able to decrease the bacterial burden in the lungs and spleens of infected mice over one month of treatment to a level lower than that of RHZ (Fig. 2). DMD and

STZ are suitable candidates for the treatment of TB as they have low toxicity in preclinical species and are well-tolerated in humans (28, 33). However, further studies on their potential cytotoxicity when combined with MCZ should be performed before use in humans, as increased HepG2 cytotoxicity was observed with this combination (Table 4).

In conclusion, MCZ enhances the activity of several potential partner drugs in vitro and in vivo. According to the results obtained in this study, BDQ, CLO, DMD and STZ are indeed good candidates to elaborate a new MCZ-containing regimen active against all forms of TB. Moreover, boosting MCZ activity by the addition of CLM is also of interest but further studies will be needed to determine the efficacy of this specific combination in vivo.

249 MATERIALS AND METHODS

Strain and culture methods.

M. tuberculosis H37Rv (Institut Pasteur) was used for the in vitro assessment of drug activity against replicating H37Rv (9). Bacteria were cultured in 7H9 complete (Middlebrook 7H9

[Difco] broth supplemented with 10% albumin-dextrose-catalase (ADC) enrichment, 0.2% glycerol, 0.05% Tween 80) at 37oC when needed.

Drug preparation.

Amikacin (AMK, Sigma), bedaquiline (BDQ, Janssen Pharmaceutica NV), clofazimine

(CLO, Sigma), clomiphene (CLM, Sigma), clarithromycin (CLR, Sigma), D-cycloserine (DCS,

Sigma), delamanid (DMD, Otsuka Co.), ethambutol (EMB, Sigma), ethionamide (ETO, Sigma), isoniazid (INH, Sigma), levofloxacin (LVX, Sigma), linezolid (LZD, AstraZeneca), lansoprazol sulfide (LPZs, Toronto Research Chemicals), meropenem (MEM, Sigma), moxifloxacin (MFX,

Sigma), para-aminosalicylic acid (PAS, Sigma), MCZ HCl salt (called MCZ throughout, was obtained from iM4TB), pyrazinamide (PZA, Fluka), rifampicin (RIF, Sigma), sutezolid (STZ,

Sequella) and tamoxifen (TMX, Sigma) were dissolved in DMSO to obtain stock concentrations of 10 mg/mL. Aliquots were stored at -20oC. When needed, the drugs were diluted in DMSO to the desired concentration. Stocks were thawed a maximum of 5 times to avoid potential loss of drug potency.

250 Antimycobacterial assays.

Drugs were tested against M. tuberculosis H37Rv strain using the resazurin microtiter assay

(REMA) in 96-well plates as previously described (15). Briefly, a mid-logarithmic phase culture of H37Rv (OD600nm approx. 0.5) was diluted in 7H9 complete medium to an OD600nm of 0.0001

(approximately 3X104 CFU/mL)%DFWHULD ȝ/ were then dispensed in transparent flat-bottom

96 well plates. Two-fold serial dilutions of each drug (resuspended in DMSO) were then prepared.

On each plate, controls without drug and media alone were included. Plates were incubated for six days at 37°C before addition of resazurin (0.025% [wt/vol] to 1/10 of the well volume). After overnight incubation, the fluorescence of the resazurin metabolite, resorufin, was determined with excitation at 560 nm and emission at 590 nm, measured using a TECAN Infinite M200 microplate reader. The MIC was calculated using GraphPad Prism software (version 7) and the Gompertz equation for MIC determination. All drugs were tested at least in duplicate.

Checkerboard assay.

To assess if compounds in two-drug combinations acted synergistically, antagonistically, or indifferently, a checkerboard assay was employed using REMA. The assay was performed in 96 well plates as sketched in Fig. S1. Fractional inhibitory concentrations (FICs) were calculated using the following formula:

FIC (X + Y) = [MIC of compound X in combination with compound Y] / [MIC of X alone].

The fractional inhibitory index (FICI) was calculated as FICI of compound X + FICI of

FRPSRXQG

Ȉ),&,YDOXHVDUHLQGLFDWLYHRIDQWDJRQLVPThe MIC in this assay is defined by the lowest

251 concentration that inhibits at least 90% of the bacterial growth (MIC90). Values in between correspond to additivity (indifferent). FICIm and FICIM were also determined and reported as the lowest FICI (FICIm) and the highest FICI (FICIM) observed out of all assays performed. Also, the mean FICI was calculated for all drug combinations between the 0.5X MIC and 0.0625X MIC. In

RI DOO FICI obtained between the 0.5X MIC and ,&,(گ UHSUHVHQWV WKH PHDQ ,&,(گ WKH WDEOH

0.0625X MIC of MCZ HCl. The experiments were performed in biological duplicates.

When mentioned, the interaction between MCZ and drugs was also assessed using the luminescent H37Rv-pEG200 strain, as previously described (34). A mid-logarithmic phase culture of H37Rv-PEG200 (OD600nm approx. 0.5) was diluted in 7H9 complete to an OD600nm of 0.001

(approximately 3X105 CFU/mL). Luminescence was measured by using a TECAN Infinite M200 microplate reader after 6 days incubation at 37oC. The MIC was calculated using GraphPad Prism software (version 7) and the Gompertz equation for MIC determination. All combinations were tested at least in duplicate.

CFU enumeration after drug exposure in vitro.

To assess the activity of anti-TB drugs on bacterial viability, H37Rv was diluted to an

OD600nm of 0.0001 and exposed to the drugs using the same protocol as for REMA. After seven days LQFXEDWLRQȝ/RI bacteria were 10-fold serially-diluted in PBS containing 0.05% Tween80 and plated on 7H10 agar plates containing 10% OADC (Gibco). CFUs were counted after four weeks incubation at 37oC. D0 (the day the bacteria were exposed to the drugs) and D7 (after seven days incubation with the drugs) were also plated as a control to determine the bacterial growth. For each experiment, a REMA assay was performed in addition to determine the MIC of the drugs tested. Significant results for drugs tested alone were reported as a decrease of at least 2 log10

252 CFU/mL compared to the mean CFU/mL of the non-treated control (NT) at D7. For combinations, significant effects were reported as a decrease of at least 2 log10 CFU/mL compared to the mean

CFU/mL of drugs tested alone.

Cytotoxicity assay.

Drugs were tested individually in duplicate for their cytotoxicity against human hepatic

HepG2 cells (ATCC® HB-8065™) by REMA (10). Drugs were prepared in DMSO to a final concentration of 100 ȝg/mL and were then added to 96-well plates containing 4.0 X104 cells/well in a final volume of 100 ȝ/ 7KH GUXJV ZHUH WKHQ VHULDOly diluted two-fold. The plates were incubated for three days at 37°C and fluorescence, after addition of resazurin, was determined using a TECAN M200 microplate reader. Cytotoxic effect was expressed as TD50 and TD99, which correspond to the toxic doses inhibiting growth of 50% and 99% of the HepG2 cells, respectively.

To assess the potential cytotoxic effect of two-drug regimens combining MCZ with another drug in HepG2 cells, a mixture of MCZ (final concentration in the plate: 50 ȝg/mL) and each of the other drugs (final concentration 50 ȝg/mL) was added to the plates containing HepG2 cells.

The mixture was then two-fold serially diluted and incubated three days at 37oC before resazurin was added to the plates.

In vivo assessment of MCZ in combination with DMD and STZ.

Female BALB/c mice (5-6 week-old) were obtained from Charles River Laboratories. Mice were subjected to a low-dose aerosol infection of M. tuberculosis H37Rv. Treatment began 4 weeks

253 after infection and compounds were administered by gavage 5 days a week for 4 weeks to infected mice, so as to determine the in vivo efficacy of single drugs and combinations.

DMD and STZ activity were first assessed in combination with MCZ at the following doses

(mg/kg): STZ: 100; MCZ 25; DMD 5; PZA, 150; RIF, 10; INH, 25. In this experiment, 5% Arabic gum in distilled water was used as the vehicle. At the end of the experiment, all mice were sacrificed and the bacterial load in the lungs and spleen was determined by plating dilutions of organ homogenates on 7H10 agar plates containing 10% OADC, cycloheximide (10 ȝJPO and ampicillin (50 ȝJPO Plates were incubated for 4 weeks at 37°C before CFU were enumerated.

CFU counts were log10 transformed before analysis as mean log10 CFU ± s.d., and compared using

Student's t-tests in Prism version 7.0 (GraphPad). Animal experiments were approved by the Swiss

Cantonal Veterinary Authority (authorization no. 3082).

254 ACKNOWLEDGMENTS

We would like to thank Janssen Pharmaceutical, Otsuka Novel Products GmbH, and

Sequella, Inc for kindly providing the BDQ, DMD, and STZ used in this study, respectively.

FUNDING INFORMATION

This work was funded by the European Community’s Seventh Framework Programme under grant agreement 260872.

REFERENCES

1. WHO | Global tuberculosis report 2017. WHO.

2. Maitra A, Bates S, Kolvekar T, Devarajan PV, Guzman JD, Bhakta S. 2015. Repurposing—a

ray of hope in tackling extensively drug resistance in tuberculosis. Int J Infect Dis 32:50–55.

3. Ramón-García S, Río RG del, Villarejo AS, Sweet GD, Cunningham F, Barros D, Ballell L,

Mendoza-Losana A, Ferrer-Bazaga S, Thompson CJ. 2016. Repurposing clinically approved

cephalosporins for tuberculosis therapy. Sci Rep 6:srep34293.

4. Pagliotto ADF, Caleffi-Ferracioli KR, Lopes MA, Baldin VP, Leite CQF, Pavan FR, Scodro

RB de L, Siqueira VLD, Cardoso RF. 2016. Anti-Mycobacterium tuberculosis activity of

antituberculosis drugs and amoxicillin/clavulanate combination. J Microbiol Immunol Infect

49:980–983.

255 5. Ramón-García S, Ng C, Anderson H, Chao JD, Zheng X, Pfeifer T, Av-Gay Y, Roberge M,

Thompson CJ. 2011. Synergistic Drug Combinations for Tuberculosis Therapy Identified by

.Antimicrob Agents Chemother 55:3861–3869 .ۅa Novel High-Throughput Screen

6. Li W, Sanchez-Hidalgo A, Jones V, Moura VCN de, North EJ, Jackson M. 2017. Synergistic

Interactions of MmpL3 Inhibitors with Antitubercular Compounds In Vitro. Antimicrob

Agents Chemother 61:e02399-16.

7. WHO | Target Regimen Profiles for TB Treatment. WHO.

8. Brecik M, Centárová I, Mukherjee R, Kolly GS, Huszár S, Bobovská A, Kilacsková E,

Mokošová V, Svetlíková Z, Šarkan M, Neres J, Korduláková J, Cole ST, Mikušová K. 2015.

DprE1 Is a Vulnerable Tuberculosis Drug Target Due to Its Cell Wall Localization. ACS

Chem Biol 10:1631–1636.

9. Zhang M, Sala C, Hartkoorn RC, Dhar N, Mendoza-Losana A, Cole ST. 2012. Streptomycin-

Starved Mycobacterium tuberculosis 18b, a Drug Discovery Tool for Latent Tuberculosis.

Antimicrob Agents Chemother 56:5782–5789.

10. Makarov V, Lechartier B, Zhang M, Neres J, van der Sar AM, Raadsen SA, Hartkoorn RC,

Ryabova OB, Vocat A, Decosterd LA, Widmer N, Buclin T, Bitter W, Andries K, Pojer F,

Dyson PJ, Cole ST. 2014. Towards a new combination therapy for tuberculosis with next

generation benzothiazinones. EMBO Mol Med 6:372–383.

11. Lechartier B, Cole ST. 2015. Mode of Action of Clofazimine and Combination Therapy with

Benzothiazinones against Mycobacterium tuberculosis. Antimicrob Agents Chemother

59:4457–4463.

256 12. Matteelli A, Carvalho AC, Dooley KE, Kritski A. 2010. TMC207: the first compound of a

new class of potent anti-tuberculosis drugs. Future Microbiol 5:849–858.

13. Yano T, Kassovska-Bratinova S, Teh JS, Winkler J, Sullivan K, Isaacs A, Schechter NM,

Rubin H. 2011. Reduction of clofazimine by mycobacterial type 2 NADH:quinone

oxidoreductase: a pathway for the generation of bactericidal levels of reactive oxygen

species. J Biol Chem 286:10276–10287.

14. Degang Y, Akama T, Hara T, Tanigawa K, Ishido Y, Gidoh M, Makino M, Ishii N, Suzuki

K. 2012. Clofazimine modulates the expression of lipid metabolism proteins in

Mycobacterium leprae-infected macrophages. PLoS Negl Trop Dis 6:e1936.

15. Palomino J-C, Martin A, Camacho M, Guerra H, Swings J, Portaels F. 2002. Resazurin

Microtiter Assay Plate: Simple and Inexpensive Method for Detection of Drug Resistance in

Mycobacterium tuberculosis. Antimicrob Agents Chemother 46:2720–2722.

16. Zumla A, Nahid P, Cole ST. 2013. Advances in the development of new tuberculosis drugs

and treatment regimens. Nat Rev Drug Discov 12:388–404.

17. Tiberi S, Payen M-C, Sotgiu G, D’Ambrosio L, Alarcon Guizado V, Alffenaar JW, Abdo

Arbex M, Caminero JA, Centis R, De Lorenzo S, Gaga M, Gualano G, Roby Arias AJ,

Scardigli A, Skrahina A, Solovic I, Sulis G, Tadolini M, Akkerman OW, Alarcon Arrascue E,

Aleska A, Avchinko V, Bonini EH, Chong Marín FA, Collahuazo López L, de Vries G, Dore

S, Kunst H, Matteelli A, Moschos C, Palmieri F, Papavasileiou A, Spanevello A, Vargas

Vasquez D, Viggiani P, White V, Zumla A, Migliori GB. 2016. Effectiveness and safety of

257 meropenem/clavulanate-containing regimens in the treatment of MDR- and XDR-TB. Eur

Respir J 47:1235–1243.

18. Van der Paardt A-FL, Akkerman OW, Gualano G, Palmieri F, Davies Forsman L, Aleksa A,

Tiberi S, de Lange WCM, Bolhuis MS, Skrahina A, van Soolingen D, Kosterink JGW,

Migliori GB, van der Werf TS, Alffenaar J-WC. 2017. Safety and tolerability of

clarithromycin in the treatment of multidrug-resistant tuberculosis. Eur Respir J 49.

19. Gopal M, Padayatchi N, Metcalfe JZ, O’Donnell MR. 2013. Systematic review of

clofazimine for the treatment of drug-resistant tuberculosis. Int J Tuberc Lung Dis Off J Int

Union Tuberc Lung Dis 17:1001–1007.

20. Migliori GB, Eker B, Richardson MD, Sotgiu G, Zellweger J-P, Skrahina A, Ortmann J,

Girardi E, Hoffmann H, Besozzi G, Bevilacqua N, Kirsten D, Centis R, Lange C, TBNET

Study Group. 2009. A retrospective TBNET assessment of linezolid safety, tolerability and

efficacy in multidrug-resistant tuberculosis. Eur Respir J 34:387–393.

21. Feng X, Zhu W, Schurig-Briccio LA, Lindert S, Shoen C, Hitchings R, Li J, Wang Y, Baig

N, Zhou T, Kim BK, Crick DC, Cynamon M, McCammon JA, Gennis RB, Oldfield E. 2015.

Antiinfectives targeting enzymes and the proton motive force. Proc Natl Acad Sci

112:E7073–E7082.

22. Li W, Upadhyay A, Fontes FL, North EJ, Wang Y, Crans DC, Grzegorzewicz AE, Jones V,

Franzblau SG, Lee RE, Crick DC, Jackson M. 2014. Novel Insights into the Mechanism of

Inhibition of MmpL3, a Target of Multiple Pharmacophores in Mycobacterium tuberculosis.

Antimicrob Agents Chemother 58:6413–6423.

258 23. Lechartier B, Hartkoorn RC, Cole ST. 2012. In vitro combination studies of benzothiazinone

lead compound BTZ043 against Mycobacterium tuberculosis. Antimicrob Agents Chemother

56:5790–5793.

24. 2006. Use of clomiphene citrate in women. Fertil Steril 86:S187–S193.

25. Farha MA, Czarny TL, Myers CL, Worrall LJ, French S, Conrady DG, Wang Y, Oldfield E,

Strynadka NCJ, Brown ED. 2015. Antagonism screen for inhibitors of bacterial cell wall

biogenesis uncovers an inhibitor of undecaprenyl diphosphate synthase. Proc Natl Acad Sci

112:11048–11053.

26. Jang WS, Kim S, Podder B, Jyoti MA, Nam K-W, Lee B-E, Song H-Y. 2015. Anti-

Mycobacterial Activity of Tamoxifen Against Drug-Resistant and Intra-Macrophage

Mycobacterium tuberculosis. J Microbiol Biotechnol 25:946–950.

27. Kalia NP, Hasenoehrl EJ, Ab Rahman NB, Koh VH, Ang MLT, Sajorda DR, Hards K,

Grüber G, Alonso S, Cook GM, Berney M, Pethe K. 2017. Exploiting the synthetic lethality

between terminal respiratory oxidases to kill Mycobacterium tuberculosis and clear host

infection. Proc Natl Acad Sci U S A 114:7426–7431.

28. Wallis RS, Dawson R, Friedrich SO, Venter A, Paige D, Zhu T, Silvia A, Gobey J, Ellery C,

Zhang Y, Eisenach K, Miller P, Diacon AH. 2014. Mycobactericidal Activity of Sutezolid

(PNU-100480) in Sputum (EBA) and Blood (WBA) of Patients with Pulmonary

Tuberculosis. PLoS ONE 9.

259 29. Sulochana S, Mitchison DA, Kubendiren G, Venkatesan P, Paramasivan CN. 2009.

Bactericidal activity of moxifloxacin on exponential and stationary phase cultures of

Mycobacterium tuberculosis. J Chemother Florence Italy 21:127–134.

30. Riccardi N, Del Puente F, Magnè F, Taramasso L, Di Biagio A. 2017. Bedaquiline: a new

hope for shorter and better anti-tuberculosis regimens. Recent Patents Anti-Infect Drug Disc.

31. Grosset J, Vernon A. 2015. A reader’s guide to the bactericidal activity of pyrazinamide and

clofazimine alone and in combinations with pretomanid and bedaquiline. Am J Respir Crit

Care Med 191:871–873.

32. Saliu OY, Crismale C, Schwander SK, Wallis RS. 2007. Bactericidal activity of OPC-67683

against drug-tolerant Mycobacterium tuberculosis. J Antimicrob Chemother 60:994–998.

33. Xavier AS, Lakshmanan M. 2014. Delamanid: A new armor in combating drug-resistant

tuberculosis. J Pharmacol Pharmacother 5:222–224.

34. Sharma S, Gelman E, Narayan C, Bhattacharjee D, Achar V, Humnabadkar V,

Balasubramanian V, Ramachandran V, Dhar N, Dinesh N. 2014. Simple and Rapid Method

To Determine Antimycobacterial Potency of Compounds by Using Autoluminescent

Mycobacterium tuberculosis. Antimicrob Agents Chemother 58:5801–5808.

260 Figures and Tables

261 Figure 1: CFU counting after exposure to MCZ in two-drug combinations. (A) NT (non-treated)

M. tuberculosis H37Rv viability on day 0 and on day 7 was assessed in the presence of INH and

RIF, (1X MIC [grey bars] and 0.25X MIC [white bars]) as controls. (B) Activity of drugs tested individually at 2X MIC (black bars), 1X MIC (grey bars), 0.5X MIC (tiled bars) and 0.25X MIC

(white bars). (C) The activity of MCZ two-drug combinations at 1X MIC (grey bars) and 0.25X

MIC (white bars) of each drug. Faint dashed line represents the limit of detection in the assay. Dark dashed lines represent the number of CFU of the non-treated bacteria at D0 and after 7 days incubation. A decrease of at least 2 log10 in the mean CFU/mL was considered significant (*) compared to the non-treated (NT) control after 7 days incubation (Fig. 1 A and 1B). In combination, a decrease of 2 log10 CFU/mL was considered significant (*) compared to the drugs alone (Fig. 1C);

262 Figure 2: Efficacy of drugs alone or in combination in a mouse model of TB. Drugs were administered by gavage at the following doses: rifampicin (R) = 10mg/kg, isoniazid (H) = 25mg/kg, pyrazinamide (Z) = 150mg/kg, MCZ (MCZ) = 25mg/kg, delamanid (DMD) = 5mg/kg, sutezolid

(STZ) = 100mg/kg. White and black columns correspond to the bacterial burden in the lungs and spleens respectively at day 0 (D0) when treatment was initiated, or day 28 (NT, no treatment; regimens as indicated) when treatment ended. Bars represent the mean ± s.d. of CFUs from five

Balb/c mice per group. *, P P 0.0002

263 Table 1: MIC of the drugs against M. tuberculosis H37Rv

Drugs 0,& ȝJP/ AMK 0.02 BDQ 0.07 CLO 0.20 CLM 24.8 CLR 14.2 DCS 2.80 DMD 0.0017 EMB 0.15 ETO 1.53 INH 0.28 LVX 0.17 LZD 0.34 LZPs 1.4 MEM 1.00 MFX 0.03 MCZ HCl 0.0002 PAS 0.014 PZA NDa RIF 0.0005 STZ 0.20 TMX 14.22 Abbreviations: AMK, Amikacin; BDQ, Bedaquiline; CLO, clofazimine; CLM, clomiphene; CLR, clarithromycin; DCS, D-cycloserine; DMD, delamanid; EMB, ethambutol; ETO, ethionamide; INH, isoniazid; LVX, levofloxacin; LZD, linezolid; LPZS, lansoprazole sulfide; MEM, meropenem; MFX, moxifloxacin; PAS, para-aminosalicylic acid; PZA, pyrazinamide; RIF, rifampicin; STZ, sutezolid; TMX, tamoxifen. aND, not detectable in the conditions tested

264 Table 2: Combinatory effect of MCZ with first- and second-line drugs against M. tuberculosis

H37Rv in vitro (checkerboard assay)

Combinations FICIm FICIM Mean FICI Mean FICI Combinatory effect a b c d (0.25X MIC) (FICIm/FICIM) MCZ+AMK 1.125 2.000 1.250 1.469 I/I MCZ+DCS 1.250 3.000 1.125 1.844 I/I MCZ+EMB 1.125 2.000 1.250 1.844 I/I MCZ+ETO 1.125 1.500 1.250 1.344 I/I MCZ+INH 1.063 1.500 1.250 1.236 I/I MCZ+LVX 1.125 2.000 1.125 1.469 I/I MCZ+MFX 1.125 2.000 1.250 1.469 I/I MCZ+PAS 1.125 2.000 1.250 1.469 I/I MCZ+ RIF 0.750 1.125 1.250 0.922 I/I a Represents the lowest FICI observed in this assay b Represents the highest FICI observed in this assay c Calculated at 0.25X MIC of the first drug indicated in the combination. d Combinatory effect: I, indifferent and S, synergism

265 Table 3: Combinatory effect of MCZ with repurposed or new anti-TB drugs against M. tuberculosis H37Rv in vitro (checkerboard assay)

b c Combinations FICIm FICIM Mean FICI Mean Combinatory effect d e (0.25X MIC) FICI (FICIm/FICIM) MCZ+BDQ 0.375 1.250 0.375 0.719 S/I MCZ+CLM 0.250 1.063 0.250 0.633 S/I MCZ+CLO 0.375 3.000 0.375 0.719 S/I MCZ+CLR 1.033 1.250 1.250 1.063 I/I MCZ+DMD 1.125 2.500 1.125 1.469 I/I MCZ+LPZs 1.125 2.500 1.250 1.469 I/I MCZ+LZD 1.125 2.000 1.250 1.328 I/I MCZ+MEM 1.063 1.500 1.250 1.146 I/I MCZ+STZ 1.063 2.500 1.250 1.146 I/I MCZ+TMXa 0.625 1.063 0.750 0.935 I/I a Combination was assessed using the H37Rv-lux strains as previously reported. b Represents the lowest FICI observed in this assay c Represents the highest FICI observed in this assay d Calculated at 0.25X MIC of the first drug indicated in the combination. e Combinatory effect: I, indifferent and S, synergism Combinations in bold are synergistic

266 Table 4: Cytotoxicity assay of drug combinations in human HepG2 cells

Drugs/combinations TD50 TD99 (ȝg/mL)a (ȝg/mL)b BDQ 9.68 40.47 CLO 0.91 2.82 DMD 52.87 56.81 MCZ 51.16 56.22 STZ 51.32 71.61 BDQ+CLO 0.66 1.99 DMD+BDQ 5.32 14.24 DMD+CLO 0.50 1.07 DMD+STZ 24.85 36.74 MCZ+BDQ 11.19 26.02 MCZ+CLO 1.08 2.61 MCZ+DMD 6.76 21.15 MCZ+STZ 34.50 55.01 MCZ+CLO+BDQ 0.77 2.04 MCZ+DMD+BDQ 3.68 9.01 MCZ+DMD+CLO 0.58 1.77 MCZ+STZ+DMD 5.61 17.23 a Toxic dose that inhibits 50% of the cell viability b Toxic dose that inhibits 99% of the cell viability

267 Supplemental material

Table S1: M. tuberculosis H37Rv viability after treatment with two-drug combinations (Figure 1)

MIC of each drug 0.25X MIC of each drug log10 CFU/mL s.d. log10 CFU/mL s.d. NT D0 3.950 0.123 3.950 0.123 NT D7 6.996 0.627 6.996 0.627 INH 7.444 0.209 7.496 0.235 RIF 4.543 1.100 7.378 0.214 PZA 7.159 0.064 7.231 0.327 INH+RIF 3.778 0.988 7.224 0.110 INH+RIF+PZA 4.863 1.544 6.628 0.462 MCZ 3.701 1.865 7.235 0.675 BDQ 4.705 0.817 7.289 0.588 MCZ+BDQ 2.531 0.213 6.477 0.000 CLO 3.690 1.748 7.366 0.157 MCZ+CLO 2.000 0.000 5.334 0.223 DMD 3.200 1.137 6.804 0.390 MCZ+DMD 1.699 0.000 7.054 0.213 MFX 4.690 0.301 6.835 0.503 MCZ+MFX 3.223 0.380 6.977 0.281 STZ 5.955 0.297 6.780 0.386 MCZ+STZ 2.523 0.064 6.889 0.269 BDQ+DMD 3.914 0.015 7.386 0.257 BDQ+STZ 4.699 0.000 7.521 0.311 CLO+BDQ 1.810 1.404 5.352 1.024 CLO+DMD 3.727 0.225 6.163 0.019 CLO+STZ 4.006 0.211 5.921 0.018 DMD+STZ 4.651 0.069 7.000 0.563 MFZ+BDQ 4.182 0.106 6.393 0.859 MFZ+CLO 5.175 0.041 6.275 0.164

268 Figure S1: Schematic representation of checkerboard plates as they were prepared for this study.

Gradient drug A

A Bacteria ~ MIC Media drug A B Bacteria 4 x MIC dug B Media

C Bacteria 2 x MIC drug B Media BGradient drug D Bacteria 1 x MIC drug B Media E Bacteria ½ x MIC drug B Media F Bacteria ¼ x MIC drug B Media

G Bacteria 1/8 x MIC drug B Media

H Bacteria 1/16 x MIC drug B Media

269 A

C.glutamicum 1 ------MSEFQVP M.tuberculosis 1 VARDARKRTSSNFPQLPPAPDDYPTFPDTSTWPVVFPELPAAPYGGPCRPPQHTSKAAAP S.aureus 1 ------MFKKLINK------KNTINNYNE E.coli 1 ------MMLSA------TQPLSE

C.glutamicum 8 EIPAQFLPKHIALVMDGNGRWATERGMKRTEGHKRGEAVLLDVVDACIELGVPYLSAYAF M.tuberculosis 61 RIPADRLPNHVAIVMDGNGRWATQRGLARTEGHKMGEAVVIDIACGAIELGIKWLSLYAF S.aureus 18 ELDSSNIPEHIAIIMDGNGRWAKKRKMPRIKGHYEGMQTIKKITRIASDIGVKYLTLYAF E.coli 12 K-LPAHGCRHVAIIMDGNGRWAKKQGKIRAFGHKAGAKSVRRAVSFAANNGIEALTLYAF

C.glutamicum 68 STENWRRSTDEVRFLMGFNRDVLRRQRDDLHEKGVRVRWVGRRPRLWRSVIRELETAEEL M.tuberculosis 121 STENWKRSPEEVRFLMGFNRDVVRRRRDTLKKLGVRIRWVGSRPRLWRSVINELAVAEEM S.aureus 78 STENWSRPESEVNYIMNLPVNFLKTFLPELIEKNVKVETIGFTDKLPKSTIEAINNAKEK E.coli 71 SSENWNRPAQEVSALMELFVWALDSEVKSLHRHNVRLRIIGDTSRFNSRLQERIRKSEAL

C.glutamicum 128 TKDNTTMTLAMCVNYGGRAEIIDAARDIARLAAEGKLRPEQITEKTFPNFLDEPDMPDVD M.tuberculosis 181 TKSNDVITINYCVNYGGRTEITEATREIAREVAAGRLNPERITESTIARHLQRPDIPDVD S.aureus 138 TANNTGLKLIFAINYGGRAELVHSIKNMFDELHQQGLNSDIIDETYINNHLMTKDYPDPE E.coli 131 TAGNTGLTLNIAANYGGRWDIVQGVRQLAEKVQQGNLQPDQIDEEMLNQHVCMHELAPVD

C.glutamicum 188 LFLRPSGEKRTSNFLLWQSAYAEMVYQDKLFPDFTQQDLYDAVLEYAKRDRRFGSA---- M.tuberculosis 241 LFLRTSGEQRSSNFMLWQAAYAEYIFQDKLWPDYDRRDLWAACEEYASRTRRFGSA---- S.aureus 198 LLIRTSGEQRISNFLIWQVSYSEFIFNQKLWPDFDEDELIKCIKIYQSRQRRFGGLSEE- E.coli 191 LVIRTGGEHRISNFLLWQIAYAELYFTDVLWPDFDEQDFEGALNAFANRERRFGGTEPGD

C.glutamicum --- M.tuberculosis --- S.aureus --- E.coli 251 ETA

C. glutamicum M. tuberculosis S. aureus E. coli C. glutamicum 100.00 64.61 37.86 38.17 M. tuberculosis 64.61 100.00 37.94 39.02 S. aureus 37.86 37.94 100.00 39.76 E. coli 38.17 39.02 39.76 100.00

Figure S2: Protein sequence alignments of UppS. Protein sequences from C. glutamicum ATCC 13032 (NCgl2203), M. tuberculosis H37Rv (Rv2361c), Staphylococcus aureus N315 (SA1103) and Escherichia coli K12 substr. MG1655 (b0174) were aligned using CLUSTAL Omega (A). Amino acids in red were reported to interact with clomiphene (PDB: 5CQJ). Identities (%) between amino acid sequences (B) were calculated using CLUSTAL 2.1.

270 Acknowledgements I would like to express my heartfelt gratitude to the UPCOL members, both past and present, for creating a stimulating, dynamic, and fun environment, for looking out for each other, and for making this experience a truly memorable and meaningful one. I am extremely thankful, first and foremost, to Prof. Stewart Cole, for adopting me when I was lab-less and giving me the opportunity to be part of an incredible team, and for his guidance, supervision, and encouragement throughout these years. Stewart is an inspirational figure and has impressed much on me, including his broad knowledge, sharp observational skills, and humour. I would also like to thank Claudia for her initiation into the P3 lab, for being a great teacher and her patience in teaching, and for always having some time for discussions; Jérémie for his many (wacky) ideas and scientific input on projects, for being always up for scientific and non-scientific discussions, for his motivation and encouragement, for polishing manuscripts and the draft of this thesis; Andréanne for her guidance and collaboration on drug discovery projects, for many fruitful discussions based on her microbiology and drug resistance experience, and for making long P3 times more fun and offering help in hard times (e.g. gavaging mice, 32 REMA plates); Andrej for always having his door open and willing to take time to explain some bioinformatics (and inappropriate jokes); the lab technicians, Anthony, for all the huge initial screens of compounds and help with the mice, Stef, for ensuring the smooth running of the lab, and for always being really kind, Philippe for helping to run the lab and sharing interesting trivia facts; Jan who took time to demonstrate the tricky business of M. tuberculosis recombineering; Rita for discussions related to drug discovery and other topics, and for sharing nice moments at MM4TB meetings together with Monica; and to the PhD students, Ye, Nina, Charlotte, Paloma, Raphael, and Sofia, for always being ready to help with experiments, sharing tips and advice, as well as the ups and downs of PhD life. A special mention also to Gaëlle, for her introduction to DprE1 at the beginning; to Benoit, for an engaging project to take over from; to Jean-Yves, for PBTZ interactions and a few fresh air walks; and to Cécile and Suzanne who have enthusiastically and whole-heartedly provided administrative support, and listening ears. I would also like to thank the chemists, Prof. Karl-Heinz Altmann and his team, Maryline, Tobias, and Patrick, for very fruitful collaborations, for the engaging discussions during our bi-annual meetings, and for generating and synthesising interesting molecules to work with; and to our other collaborators with whom I have had the pleasure of interacting with, Dirk and Prof. Adrie Steyn, Prof. Kevin Pethe, and Prof. Mary Jackson, through email correspondence and/or meetings at conferences, and to Dr. Helena Boshoff for taking time in providing constructive input on the PB project. I am also thankful to members of the labs of the Global Health Institute, particularly the neighbouring Blokesch, McKinney and Lemaitre labs, for the floor interactions in and out of the labs, and for always being willing to lend a helping hand in dire situations, such as when there is a shortage of material; and to my mentor, Prof. Bart Deplancke, for his annual words of encouragement. I am also very grateful towards members of the now-defunct UPMOORE lab, to Prof. Darren Moore for giving me the opportunity to work in his lab for my Master thesis and for accepting me thereafter as a PhD candidate, for his close guidance and supervision; to Elpida for supervising my Master project and for teaching me many experimental techniques patiently, Guillaume, Roger, Alessandra, Agata, and Liliane, for all the experimental tips and tricks,and scientific advice; and to Duygu, Meghna, and Ece, for sharing the life of Master students. I also greatly appreciate Prof. Melanie Blokesch, Prof. Karl-Heinz Altmann, Prof. Peter Sander, and Prof. Mary Jackson, for taking time out of their schedules to participate as jury members of the committee. Last, but certainly not the least, I am extremely thankful for the enriching interactions, inspiration, and support of those whom I have met over the course of my time in Switzerland, for the friendships, for the good times, and for the love and support of those dearest to me.

271

&DUROLQH 6KL

./ - *! $ ) Ѷ $*( $' )"$) -$)" - ǓǑǒǑ Ȏ +/ ǓǑǒǓ )$1 -.$/4 *! ++'$ $ ) . # ) -()4 ғ 3#)" ./0 )/ / Ѷ 2$/5 -') җ*1 спрр Ҋ 0" спрсҘ # '*- 2$/# *)*0-.Ѷ $* )"$) -$)" 0" ǓǑǑǖ Ȏ 4 ǓǑǑǚ /$*)' )$1 -.$/4 *! $)"+*- $)"+*- ғ 3#)" ./0 )/ / #)$' )$1 -.$/4 *! ƕ)$#Ѷ -()4 җ/ сппц Ҋ 0" сппчҘ

3+ -$ ) #ѵ - . -# 0" ǓǑǒǕ Ȏ +- . )/ *-/*-4 *! $-*$' /#*" ) .$. җ ҘѶ '*' '/# )./$/0/ Ѷ 2$/5 -') ғ ) ()" -0" $.*1 -4 +-*% /. !*- 0 -0'*.$. ғ " ) -/$*) ) *+/$($./$*) *! с ) 2 *(+*0) .Ѷ $)'0 $)" /-" / $ )/$!$/$*) ) ( #)$.( *! /$*) ғ *)/-$0/ /* '$)$' 1 '*+( )/ .0++*-/ *! 0 -0'*.$. -0" ) $ / *5$)*) 4 '0$ /$)" ( #)$.(. *! - .$./) Ѷ " ) -/$*) *! &0+ . -$ .Ѷ #-/ -$./$*) *! -0" $)/ -/$*). #$)" ..$./)/ ǓǑǒǕ Ȏ ) ǓǑǒǗ )/ "-/ *-/*-4 *0-. !*- $! $ ) .Ѷ # '*- *0-. Ѷ 2$/5 -') ғ 0+ -1$. "-*0+. *! ./0 )/. $) '*-/*-4 . //$)" ғ (*)./-/$*) ) 3+')/$*) *! $*'*"$' '*-/*-4 / #)$,0 .Ѷ .0+ -1$.$*) $) 3 0/$*) *! 3+ -$Ҋ ( )/.Ѷ $.0..$*) ) )'4.$. *! - .0'/. ғ - $)" *! ' )*/ **&. ) +-*% / - +*-/. . -# ..$./)/ / ǓǑǒǓ Ȏ - ǓǑǒǔ *-/*-4 *! 0-* " ) -/$1 . -# җ ҘѶ -$) $) )./$/0/ Ѷ 2$/5 -') ғ )1 ./$"/$*) *! /# -*' *! с $) ($-* - "0'/$*) 4 / /$)" $Ȃ - )/$' 3+- ..$*) *! ($-*. $) -&$).*)ҁ. $. . .// ғ '0$ /$*) *! ''0'- ) (*' 0'- ( #$.(. *! сҊ'$)& җ 0$) Ҋ-$# - + / &$). сҘ -&$).*)ҁ. $. . 0.$)" 1$-'Ҋ. " ) /-).! - ( /#* $) $) 1$/-* ) $) 1$1* (* '. ./ - /# .$. - . -# *1 ǓǑǒǒ Ȏ 0" ǓǑǒǓ *-/*-4 *! 0-* " ) -/$1 . -# җ ҘѶ -$) $) )./$/0/ Ѷ 2$/5 -') ғ ./'$.# ) 0-*)' /*3$$/4 ..4. $) +-$(-4 ) 0-*)' 0'/0- . . *) ) 0-*)' /#Ѷ ) 0-$/ ' )"/# ) -)#$)" *(+' 3$/4 0.$)" #*'' )'4.$. ғ 1 '*+( )/ *! +-$(-4 ) 0-*)' 0'/0- (* ' *! сҊ'$)& -&$).*)ҁ. $. . #$)" 0./$/0/ 0' ǓǑǑǚ Ȏ ǓǑǒǑ Ȃ' . )./$/0/$*) $)"+*- ғ *) 0/ ' ..*). $) #4.$. ) $*'*"4 / /# . *) -4 .#**' ' 1 ' ғ - +- ./0 )/. !*- ) Ҋ*!Ҋ4 - 3(.Ѷ +-*1$ 3/- *#$)" . ..$*). # '*- -*% / + ǓǑǑǙ Ȏ 4 ǓǑǑǚ )*Ҋ$*)'4/$. *-/*-4Ѷ /$*)' )$1 -.$/4 *! $)"+*- $)"+*- ғ +/$($./$*) *! +*'4( - '4 - +-* 0/$*) !*- '4 -Ҋ4Ҋ'4 - )+.0'/$*) *! +)- /$ ''. ғ '$ /$*) *! *(+*) )/. ) +- . ) *! +*'4( - '4 -. 0.$)" Ѷ .+ /-*.*+$ ) $*+#4.$' ( /#* .

273 #)$' &$''. $-*$*'*"4Ѷ (*' 0'- $*'*"4Ѷ $*# ($./-4Ѷ 2*-&$)" 2$/# '' '$) . ) +-$(-4 ) 0-*)' *-/*-4 .&$''. 0'/0- .Ѷ - *($))/ +0-$!$/$*) ) 3+- ..$*) *! +-*/ $). ) )*1$-0.Ѷ $((0)*4/*# ($./-4Ѷ !'0*- . )/ ) *)!*' $("$)" -$)$)". ) -/$!$/$*). *-&$)" $) $*.! /4 р /* т '*-/*-4Ѷ )$(' 3+ -$( )//$*) *(+0/ - .&$''. $-*.*ȅ Ȃ$ җ*- Ѷ 3 'Ѷ *2 -+*$)/ҘѶ * #*/*.#*+Ѷ ''0./-/*-Ѷ / 3

0'$/$*). Ç .$& Ѷ ))) Ѷ ** ҊѶ $& () Ѷ '0. - Ѷ ''#- Ѷ '/ - Ѷ )*// Ѷ 2.*) Ѷ 2.*) Ѷ **- ѵ спруѵ *) $/$*)' 3+- ..$*) *! -&$).*)ȅ. $. . Ȏ- '/ ǒǕǕǒ Ǔ $) ($ -$) *+($) -"$ ) 0-*). *! ($ 0. . )0' - )*-('$/$ . 2$/#*0/ ) 0-* " ) -/$*)ǹ 0-*$*' $. црѷтуфҌтфчѵ

Ç ** ҊѶ #-/$ - Ѷ *''4 Ѷ *4Ҋť//" - Ѷ - . Ѷ 4)$& - Ѷ 0+$ ) Ѷ ' Ѷ $/*) Ѷ *' ѵ спрхѵ #-Ȏ / -$5/$*)*!+-ǒȎ $/ )5*/#$5$)*) .$./) $)4*/ -$0(/0 -0'*.$.ǹ )/$($-* " )/. # (*/# - хпѷхуфрҌхуфшѵ

Ç $/*) Ѷ ** ҊѶ *' ѵспрцѵ/-0/0-' ./0 $ . *! 4*/ -$0( /0 -0'*.$. +-ǒ $)/ -/$)" 2$/# $/. $)#$$/*-.ǹ -0" $.*1 * 4 ссѷфсхҌфттѵ

Ç $/*) Ѷ */ Ѷ 0+$ ) Ѷ ** ҊѶ $*1 Ѷ &-*1 Ѷ *' ѵ спрчѵ /-0/0- Ȏ. -0" .$") ) #-Ȏ / -$5/$*) *! .0'!*)4'Ȏ+$+ -5$) )5*/#$5$)*) $)#$$/*-. *! +-ǒ !-*( 4*/ -$0( /0 -0'*.$.ǹ )/$($-* " )/. # (*/# - ѵппхчрҊрчѵ

Ç 0+$ ) Ѷ */ Ѷ ** ҊѶ '// . Ѷ $''*) ҊѶ &-*1 Ѷ *' ѵ спрчѵ ) *+/$($5 &"-*0) - "$( ) !*- /- /( )/ *! /$1 /0 -0'*.$. 2$/# /# ) 3/Ȏ" ) -/$*) )5*/#$5$)*) *5$)*) ȗǒǗǚȘ )/$($-* " )/. # (*/# - ѵ

Ç ** ҊѶ 0+$ ) Ѷ $ )' Ѷ */ Ѷ )%& Ѷ *(( - Ѷ (+- #/ Ѷ / 4) Ѷ /# Ѷ $/*) Ѷ '/()) Ҋ Ѷ *' ѵ спрчѵ -4'1$)4'+$+ -5$) ($ .Ǻ ) 2 '.. *! +*/ )/ $)#$$/*-. /-" /$)" - *! 4*/ -$0( /0 -0'*.$.ǹ ($*ѵ *)! - ) -* $)". Ç *- $$) . !*- 0 -0'*.$. җуҘ *).*-/$0( /$)" җ-$.Ѷ -) Ѷ сшҊтп/# 0) спрфҘѵ ҃#-/ -$5Ҋ /$*) *! +-рҊ $/ )5*/#$5$)*) .$./) $) 4*/ -$0( /0 -0'*.$.҃ѵ -' +- . )//$*)ǹ

Ç 0 -0'*.$. спрхѷ )/ - $.$+'$)-4 . -# *) 0 -0'*.$. ) /#*" )$ 4*/ -$ җ-$.Ѷ -) Ѷ ршҊст/# +/ ( - спрхҘѵ ** Ҋ / 'ǹѶ #-/ -$5/$*) *! +-рҊ $/ )5*/#$5$)*) .$./) $) 4Ҋ */ -$0( /0 -0'*.$.ѵ *./ - +- . )//$*)ǹ

Ç *- *) . -# *)! - ) ѷ 0 -0'*.$. -0" $.*1 -4 ) 1 '*+( )/ җ 0Ѷ /'4Ѷ сфҊтп/# 0) спрцҘѵ ** Ҋ / 'ǹѶ #-/ -$5/$*) *! +-рҊ $/ )5*/#$5$)*) .$./) $) 4*/ -$0( /0 -0'*.$.ѵ *./ - +- . )//$*)ǹ -)/.Ѷ #$ 1 ( )/. о 2- .

*0)" $ )/$./. /$)" -)/- -1 ' ''*2.#$+ !*- 0 -0*.$. спрх 0) ǓǑǒǗ *)! - ) с) ./ -' - . )//$*)- у *).*-/$ /$)" 0) ǓǑǒǖ *# *)/$) )/.- *-")$. 4 *# ѵ р *0/ *! рпп +-/$$+)/. #'' )"$)" *0) -$ . $) 0" ǓǑǒǕ .$ ) Ѷ -/ о $))*1/$*) $)"+*- *' ) -"- 0/ #*'-.#$+- )& 0" ǓǑǑǖ Ȏ 4 ǓǑǑǚ 2- !*- /0 4 -* - /$*)' )$1 -.$/4 *! $)"+*- / ǓǑǑǘ Ȏ 0" ǓǑǑǙ $+'*( $) $)* -!*-() - ..*$/ *- *! *4' #**' *! 0.$ җ$+Ҙ 0" ǓǑǑǑ )"0" .

)"'$.# /$1 ) -$) '0 )/ -() сѶ *-&$)" +-*!$$ )4 - )# р

274

Source: Stop TB Partnership